A method for assisted puncture soft tissue viscoelasticity testing based on multi-curvature sample optimization

Abstract

The present invention discloses a method for testing the viscoelasticity of soft tissues for auxiliary puncture based on multi-curvature sample optimization, including 1) performing a puncture experiment to obtain the multi-curvature characteristics of a pre-bent flexible probe; 2) establishing a deflection spring model based on a fitting deflection curve; 3) establishing a quasi-static equation based on the spring model, and solving the equation through a fourth-order ordinary differential equation; 4) establishing a small batch of samples based on the multi-curvature characteristics, and optimizing the pseudo-spring coefficient of the soft tissue through a small batch gradient descent method. The present invention obtains the viscoelastic characteristics of different soft tissues and copes with the deflection error under different puncture environments by fitting a spring model based on the multi-curvature puncture characteristics and establishing a quasi-static model based on the Euler-Bernoulli beam theory; optimizes the coefficients by a small batch gradient descent method based on the multi-curvature characteristics, performs sample optimization on the coefficients obtained under single curvature fitting, obtains the coefficient value with the minimum comprehensive error under the multi-curvature characteristics, and reduces the comprehensive deflection error in the puncture modeling of the multi-curvature probe scheme.

Description

Technical Field

[0001] The invention relates to a soft tissue viscoelasticity testing method, in particular to an auxiliary puncture soft tissue viscoelasticity testing method based on multi-curvature sample optimization. Background Art

[0002] When a flexible needle punctures soft tissue and the tip is subjected to force, the needle shaft will deflect, thereby producing an arc trajectory that can avoid obstacles. Commonly used soft tissues include silicone, agar, and gel, etc., which are all elastomers and exhibit a certain degree of viscoelasticity. Therefore, the process of flexible needle puncturing soft tissue is complicated. When the needle shaft bends, it deforms the tissue and is subjected to its elastic pressure at the same time. It is a coupled process. The viscoelastic properties of soft tissue affect the deflection of the bending needle shaft, and active needle steering schemes such as pre-bent flexible probes can obtain multi-curvature puncture trajectories. Obtaining its precise multi-curvature characteristics is closely related to the deflection of the needle shaft. In order to obtain a more accurate deflection model, the following research on auxiliary flexible needle puncture soft tissue modeling was conducted.

[0003] In 2013, Gao Dedong et al. from Qinghai University proposed a real-time measurement method for soft tissue deformation during puncture (authorization announcement number: CN104720812B). The Kelvin viscoelastic model formula was used to obtain the relationship between soft tissue deformation and soft tissue extrusion pressure based on the elastic coefficient and viscosity coefficient of the soft tissue by using Laplace transform; the real-time deformation was obtained by real-time monitoring of the change in extrusion pressure. In 2016, Li Chenggang et al. from Nanjing University of Aeronautics and Astronautics described a modeling method for contact force of multi-plane conical needles (authorization announcement number: CN105760681B). The contact modeling of irregular multi-plane conical needles was studied in the contact deformation stage, which has the effect of optimizing the needle tip design for puncture in different environments. However, due to the influence of tissue viscoelasticity during the puncture process, the deflection and force of the needle rod are also difficult to model.

[0004] However, it is difficult to accurately characterize the elastic coefficient and viscosity coefficient of soft tissues produced when studying contact and force. In order to solve the influence of soft tissue viscoelasticity on deflection, in 2017, Zhao Yanjiang and others from Harbin University of Science and Technology proposed a method for estimating the friction coefficient of soft tissue puncture by flexible needles (application publication number: CN107655824A). The friction coefficient obtained by moving the flexible needle at a constant speed includes the influence of the viscosity coefficient of soft tissue on the friction of the needle rod, which is convenient for establishing a more accurate mechanical model. Zhao Yanjiang and others also proposed a mechanical and bending deformation modeling method for soft tissue puncture by flexible needles (authorization announcement number: CN107808698B), which decomposes the needle rod friction force generally believed to be exerted on the flexible needle body into clamping friction force and tissue resistance, and considers the influence of tissue elasticity and viscosity, and conducts a complete modeling. However, the use of large deformation theory tends to ignore the influence of soft tissue viscoelasticity on needle rod deflection.

[0005] In summary, to solve the problem of modeling accuracy of assisted flexible needle puncture of soft tissue, it is necessary to consider the influence of coupled deflection of flexible needle puncture of soft tissue and the influence of different viscoelastic properties of soft tissue. To solve the above problems, the present invention combines the multi-curvature trajectory obtained by experiment, the spring model of the fitting curve and the quasi-static model to obtain the viscoelastic characteristics of soft tissue under real deflection conditions, and adopts a small batch gradient descent method based on multi-curvature feature sample optimization to obtain the viscoelastic coefficient of assisted puncture soft tissue under actual conditions, and proposes a viscoelastic test method for assisted puncture soft tissue based on multi-curvature sample optimization. Summary of the invention

[0006] The purpose of the present invention is to overcome the defects of the prior art and provide an auxiliary puncture soft tissue viscoelasticity testing method based on multi-curvature sample optimization. By fitting a spring model based on multi-curvature puncture characteristics and establishing a quasi-static model based on Euler-Bernoulli beam theory, the pseudo-spring coefficient of the spring model under a single curvature trajectory is obtained; the coefficient is optimized by a small batch gradient descent method based on multi-curvature feature sample optimization to obtain the coefficient value with the minimum comprehensive error under multi-curvature characteristics; the obtained optimized coefficient can describe the viscoelastic characteristics of different soft tissues, thereby improving the deflection modeling accuracy in different soft tissues.

[0007] The object of the present invention is achieved by: a method for auxiliary puncture soft tissue viscoelasticity testing based on multi-curvature sample optimization, comprising the following steps:

[0008] Step 1) performing a puncture experiment to obtain multi-curvature characteristics of the pre-bent flexible probe;

[0009] Step 2) establishing a deflection spring model based on the fitted deflection curve;

[0010] Step 2.1) fitting a deflection curve based on the multi-curvature characteristics of the pre-bent flexible probe deflection;

[0011] Step 2.2) establishing a deflection spring model based on the viscoelastic coefficient of soft tissue;

[0012] Step 3) establishing a quasi-static equation based on the spring model, and solving the process through a fourth-order ordinary differential equation;

[0013] Step 3.1) define the pre-bent flexible probe as a cantilever beam, and model the force on the probe based on a spring model;

[0014] Step 3.2) Use the Euler-Bernoulli theory to establish a fourth-order ordinary differential equation;

[0015] Step 3.3) Solve by four-fold integration with boundary conditions;

[0016] Step 4) establishing a small batch of samples based on the multi-curvature features, and optimizing the pseudo-spring coefficient of the soft tissue by a small batch gradient descent method;

[0017] Step 4.1) constructing the objective function of the pseudo-spring coefficient, and taking the obtained multi-curvature features as the training set;

[0018] Step 4.2) Construct an optimization formula based on the mini-batch gradient descent method.

[0019] Furthermore, the puncture experiment described in step 1) is to use a pre-bent flexible probe to puncture soft tissue in a horizontal plane, and the pre-bent flexible probe is an actively controllable flexible probe in which a pre-bent needle core and a flexible needle sleeve are coaxially arranged; multiple curvature trajectories are obtained by changing the pre-bent state, which are recorded as multi-curvature characteristics; the pre-bent state is the relative elongation of the pre-bent needle core relative to the flexible needle sleeve, and changing the pre-bent state changes the contact form with the soft tissue.

[0020] Furthermore, the fitting deflection curve in step 2) is to fit the multi-curvature features obtained in step 1) using a cubic polynomial as the needle bar bending deflection, and the needle bar bending deflection is regarded as the tissue deformation. The needle bar bending deflection ω i (x) is as shown in formula (1):

[0021] ω i (x) = a0 + a1x + a2x 2 +a3x 3 (1)

[0022] In formula (1), i is the relative elongation of the pre-bent needle core in the probe, a is the deflection fitting coefficient, A is defined as the coefficient matrix, A = [a0 a1 a2 a3], x is the probe feed amount, which is the product of the puncture speed and time.

[0023] Furthermore, the spring model established in step 2) is to approximate the process of increasing radial pressure due to compression deformation of soft tissue as a spring model. According to the spring model, the load change after the tissue is compressed is regarded as a linear process of the spring model, and the other side without compression is regarded as unchanged load. The pressure on a certain section of the flexible probe on the deflection surface is obtained as formula (2):

[0024] F t =Kω(x n ) (2)

[0025] K is the pseudo-spring coefficient of soft tissue, ω(x n ) is the bending deflection of the nth needle bar, x n is the midpoint position of the nth needle bar;

[0026] The pressure on each section of the needle rod is a load that changes with time. Therefore, the pressure on a certain section of the flexible probe is as shown in formula (3):

[0027]

[0028] Among them A i is the coefficient matrix when the needle core extension is i.

[0029] Furthermore, in step 3), the quasi-static equation based on the spring model is used to model the pre-bent flexible probe-soft tissue coupling model, and the quasi-static model of the pre-bent flexible probe is established based on the interaction between the flexible needle and the soft tissue. The pre-bent needle core and the flexible needle sleeve are modeled as the needle body, and the friction of the needle core is ignored. The flexible probe is defined as a cantilever beam, and the probe is modeled using the Euler-Bernoulli theory, and the fourth-order ordinary differential equation is obtained as shown in formula (4);

[0030]

[0031] Where x is the feed amount of the flexible probe, ω is the deflection of the flexible probe, E is the Young's modulus of the flexible probe, I is the moment of inertia of the flexible probe, α is the angle between the bevel of the flexible needle tip and the soft tissue level, and F q is the tissue pressure on the needle tip bevel, F t / l is the vertical load on the needle bar, is the slope of the needle, is the bending moment of the needle, is the shear force of the needle; the solution process is to integrate formula (4) four times respectively, and obtain the value of the integral constant through the boundary conditions each time.

[0032] Furthermore, the pseudo-spring coefficient described in step 4) is the value of the coefficient K obtained by solving the load on the needle rod based on the fourth-order ordinary differential equation; the pseudo-spring coefficient serves as a common coefficient of the influence of the soft tissue elastic coefficient and the viscosity coefficient on the deflection of the pre-bent flexible needle. In the auxiliary pre-bent flexible needle puncture soft tissue experiment, the deflection of the probe is mainly affected by the elastic properties of the soft tissue, and the influence of the viscosity coefficient on the deflection is negligible. The obtained pseudo-spring coefficient is the comprehensive actual influence coefficient.

[0033] Furthermore, the small batch gradient descent method described in step 4) is an optimization method for the coefficient K, the training set established is the deflection of the multi-curvature feature, and the objective function established is the error between the needle bar deflection based on the pseudo-spring coefficient and the corresponding multi-curvature feature. Substituting it into the initial coefficient is K intial The pseudo-spring coefficient is obtained by fitting the single curvature characteristic through the fourth-order ordinary differential equation, and the objective function J(K) is as shown in formula (5);

[0034]

[0035] Where x n is the probe feed amount of the nth needle bar, y n is the x in the multi-curvature feature n The corresponding deflection value at ω K (xn ) is the parameter K and is substituted into the quasi-static model to solve the x n The corresponding deflection value at

[0036] The small batch samples are to divide the data set into m sets, each set contains a curvature feature; the optimization method is to perform batch gradient descent on these sets one by one, and update the parameter K; until the number of iterations reaches a threshold or the cost function drops to the minimum, the update process stops, and the optimization coefficient K is obtained. final The coefficient K that minimizes the objective function J(K) is updated as shown in formula (6);

[0037]

[0038] Where (x p ,y p ) is the probe feed amount and the corresponding deflection value contained in the multi-curvature feature in the pth set; η is the learning rate.

[0039] The present invention adopts the above technical scheme, and compared with the prior art, the beneficial effects are: 1) by solving the pseudo-spring coefficient by the quasi-static equation based on the spring model, the viscoelastic characteristics of different soft tissues can be obtained to cope with the deflection error in different puncture environments; 2) a small batch of samples with multi-curvature characteristics are established, and the coefficients obtained under single curvature fitting are optimized to reduce the comprehensive deflection error in the puncture modeling of the multi-curvature probe scheme; 3) the obtained optimized coefficients can describe the viscoelastic characteristics of different soft tissues, thereby improving the deflection modeling accuracy in different soft tissues. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 Overall flow chart of the present invention.

[0041] Figure 2 Schematic diagram of the spring model of the present invention.

[0042] Figure 3 Viscoelastic coefficient optimization flow chart of the present invention. DETAILED DESCRIPTION

[0043] like Figure 1 The auxiliary puncture soft tissue viscoelasticity testing method based on multi-curvature sample optimization shown includes the following steps:

[0044] Step 1) performing a puncture experiment to obtain multi-curvature characteristics of the pre-bent flexible probe;

[0045] Step 2) establishing a deflection spring model based on the fitted deflection curve;

[0046] Step 2.1) fitting a deflection curve based on the multi-curvature characteristics of the pre-bent flexible probe deflection;

[0047] Step 2.2) establishing a deflection spring model based on the viscoelastic coefficient of soft tissue;

[0048] Step 3) establishing a quasi-static equation based on the spring model, and solving the process through a fourth-order ordinary differential equation;

[0049] Step 3.1) define the pre-bent flexible probe as a cantilever beam, and model the force on the probe based on a spring model;

[0050] Step 3.2) Use the Euler-Bernoulli theory to establish a fourth-order ordinary differential equation;

[0051] Step 3.3) Solve by four-fold integration with boundary conditions;

[0052] Step 4) establishing a small batch of samples based on the multi-curvature features, and optimizing the pseudo-spring coefficient of the soft tissue by a small batch gradient descent method;

[0053] Step 4.1) constructing the objective function of the pseudo-spring coefficient, and taking the obtained multi-curvature features as the training set;

[0054] Step 4.2) Construct an optimization formula based on the mini-batch gradient descent method.

[0055] First, the puncture experiment performed in step 1) is to puncture soft tissue in a horizontal plane using a pre-bent flexible probe, wherein the pre-bent flexible probe is an actively controllable flexible probe in which a pre-bent needle core and a flexible needle sleeve are coaxially arranged; multiple curvature trajectories are obtained by changing the pre-bent state, which are recorded as multi-curvature features; the pre-bent state is the relative elongation of the pre-bent needle core relative to the flexible needle sleeve, and changing the pre-bent state changes the contact form with the soft tissue;

[0056] Secondly, the fitting deflection curve of step 2) is to fit the multi-curvature features obtained in step 1) with a cubic polynomial as the needle bar bending deflection, and the needle bar bending deflection is regarded as the tissue deformation. The needle bar bending deflection is shown in formula (1), and ω is used i (x) represents deflection;

[0057] ω i (x) = a0 + a1x + a2x 2 +a3x 3 (1)

[0058] In formula (1), i is the relative elongation of the pre-bent needle core in the probe, a is the deflection fitting coefficient, A is defined as the coefficient matrix, A = [a0 a1 a2 a3], x is the probe feed amount, which is the product of the puncture speed and time;

[0059] like Figure 2As shown in FIG. 1 , it is a schematic diagram of the spring model of the viscoelastic test method for auxiliary puncture soft tissue based on multi-curvature sample optimization. The needle rod of the pre-bent flexible probe is evenly divided into n sections. The spring model approximates the process of increasing radial pressure due to compression deformation of soft tissue as a spring model. According to the spring model, the load change after the tissue is compressed is regarded as a linear process of the spring model, and the other side without compression is regarded as unchanged load. The pressure on a certain section of the flexible probe on the deflection surface is as shown in formula (2):

[0060] F t =Kω(x n ) (2)

[0061] K is the pseudo-spring coefficient of soft tissue, ω(x n ) is the bending deflection of the nth needle bar, x n is the midpoint position of the nth needle bar;

[0062] The pressure on each section of the needle rod is a load that changes with time. Therefore, the pressure on a certain section of the flexible probe is as shown in formula (3):

[0063]

[0064] Among them A i is the coefficient matrix when the needle core extension is i.

[0065] Then, in step 3), the quasi-static equation based on the spring model is used to model the pre-bent flexible probe-soft tissue coupling model. The quasi-static model of the pre-bent flexible probe is established based on the interaction between the flexible needle and the soft tissue. The pre-bent needle core and the flexible needle sleeve are modeled as the needle body. The friction of the needle core is ignored. The flexible probe is defined as a cantilever beam. The probe is modeled using the Euler-Bernoulli theory, and the fourth-order ordinary differential equation is obtained as shown in formula (4);

[0066]

[0067] Where x is the feed amount of the flexible probe, ω is the deflection of the flexible probe, E is the Young's modulus of the flexible probe, I is the moment of inertia of the flexible probe, α is the angle between the bevel of the flexible needle tip and the soft tissue level, and F q is the tissue pressure on the needle tip bevel, F t / l is the vertical load on the needle bar, is the slope of the needle, is the bending moment of the needle, is the shear force on the needle, which can be solved by four integrations.

[0068] When the flexible needle is punctured, the tip of the needle is deflected due to the force. The tissue in the deflection direction is compressed and changes its characteristics to produce a resisting load on the needle shaft, making the needle shaft deflect into a cantilever beam model. In order to obtain the applied load F t / l and needle tip pressure Fq The resulting deflection ω is integrated four times with respect to the feed amount x, and the first integration of equation (4) gives the shear force V(x) along the needle;

[0069]

[0070] The boundary condition for obtaining the integral constant C1 in the above formula is that the shear force on the needle tip is equal to zero, that is, x=l.

[0071]

[0072] Where A is the deflection coefficient matrix of the spring model, the constant summation term in equation (6) is C1, and the bending moment M(x) along the needle is obtained by integrating equation (6) again.

[0073]

[0074] When the boundary condition x = l, that is, the bending moment at the needle tip, the value of the integral constant C2 in the above formula is obtained. The bending moment M(x) along the needle is obtained as follows:

[0075]

[0076] The constant summation term in equation (7) is C2. Integrating equation (8) again yields the deflection slope τ(x) along the needle:

[0077]

[0078] When the boundary condition x=0, that is, the slope is 0 when the needle is fixed, the value of the integral constant C3 in the above formula is 0. The deflection slope τ(x) along the needle is obtained as follows.

[0079]

[0080] Finally, step 4) obtains the pseudo-spring coefficient of the spring model by solving the load on the needle rod based on the ordinary differential equation, and obtains the value of the coefficient K with the smallest error. During the above solution process, the shear force and torque can be obtained through the sensor at the clamping position of the probe end, and the deflection slope along the needle can be obtained from the deflection curve, and the pseudo-spring coefficient K of the soft tissue can be obtained.

[0081] like Figure 3 The figure shows the coefficient optimization process of the auxiliary puncture soft tissue viscoelastic test method based on multi-curvature sample optimization. The coefficient K is optimized by the small batch gradient descent method. The training set established is the deflection of the multi-curvature feature. The objective function established is the error between the needle rod deflection based on the pseudo-spring coefficient and the corresponding multi-curvature feature. The initial coefficient is K. intial The pseudo-spring coefficient is obtained by fitting the single curvature characteristic through the fourth-order ordinary differential equation, and the objective function J(K) is as shown in formula (11);

[0082]

[0083] Where x n is the probe feed amount of the nth needle bar, y n is the x in the multi-curvature feature n The corresponding deflection value at ω K (x n ) is the parameter K and is substituted into the quasi-static model to solve the x n The corresponding deflection value at ;

[0084] The small batch samples are divided into m sets, each of which contains a curvature feature. The optimization method is to perform batch gradient descent on these sets one by one and update the parameter K. The update process stops until the number of iterations reaches the threshold or the cost function is minimized, and the optimal coefficient K is obtained. final To minimize the coefficient K of the objective function J(K), the update of the parameter K is as shown in formula (12);

[0085]

[0086] Where (x p ,y p ) is the probe feed amount and the corresponding deflection value contained in the multi-curvature feature in the pth set; η is the learning rate.

[0087] The present invention fits a spring model based on multi-curvature puncture characteristics and establishes a quasi-static model based on the Euler-Bernoulli beam theory. The pseudo-spring coefficient of the spring model under a single curvature trajectory can be calculated, the viscoelastic characteristics of different soft tissues can be obtained, and the deflection error under different puncture environments can be dealt with. By optimizing the coefficients by a small batch gradient descent method based on multi-curvature characteristics and performing sample optimization on the coefficients obtained under single curvature fitting, the coefficient value with the minimum comprehensive error under multi-curvature characteristics can be obtained, thereby reducing the comprehensive deflection error in the puncture modeling of the multi-curvature probe scheme.

[0088] The present invention is not limited to the above-mentioned embodiments. On the basis of the technical solution disclosed in the present invention, technicians in this field can make some substitutions and deformations to some technical features therein according to the disclosed technical content without creative labor, and these substitutions and deformations are all within the protection scope of the present invention.

Claims

1. A method for assisted puncture soft tissue viscoelasticity testing based on multi-curvature sample optimization, characterized in that: The following steps are involved: Step 1) performing a puncture experiment to obtain multi-curvature characteristics of the pre-bent flexible probe; Step 2) establishing a deflection spring model based on the fitted deflection curve; Step 2.1) fitting a deflection curve based on the multi-curvature characteristics of the pre-bent flexible probe deflection; Step 2.2) establishing a deflection spring model based on the viscoelastic coefficient of soft tissue; Step 3) establishing a quasi-static equation based on the spring model, and solving the process through a fourth-order ordinary differential equation; Step 3.1) define the pre-bent flexible probe as a cantilever beam, and model the force on the probe based on a spring model; Step 3.2) Use the Euler-Bernoulli theory to establish a fourth-order ordinary differential equation; Step 3.3) Solve by four-fold integration with boundary conditions; Step 4) establishing a small batch of samples based on the multi-curvature features, and optimizing the pseudo-spring coefficient of the soft tissue by a small batch gradient descent method; Step 4.1) constructing the objective function of the pseudo-spring coefficient, and taking the obtained multi-curvature features as the training set; Step 4.2) Construct an optimization formula based on the mini-batch gradient descent method; The fitting deflection curve in step 2) is to fit the multi-curvature features obtained in step 1) using a cubic polynomial as the needle bar bending deflection, and the needle bar bending deflection is regarded as the tissue deformation. The needle bar bending deflection ω i (x) is as shown in formula (1): oh i (x)=a0+a1x+a2x 2 +a3x 3 (1) In formula (1), i is the relative elongation of the pre-bent needle core in the probe, a is the deflection fitting coefficient, A is defined as the coefficient matrix, A = [a0 a1 a2 a3], x is the probe feed amount; Step 2) Establish the deflection spring model to approximate the process of increasing radial pressure due to compression deformation of soft tissue as a spring model. According to the spring model, the load change after the tissue is compressed is regarded as a linear process of the spring model, and the other side without compression is regarded as unchanged load. The pressure on a certain section of the flexible probe on the deflection surface is obtained as formula (2): F t =Kω(x n ) (2) K is the pseudo-spring coefficient of soft tissue, ω(x n ) is the bending deflection of the nth needle bar, x n is the probe feed amount of the nth needle bar; The pressure on each section of the needle rod is a load that changes with time. Therefore, the pressure on a certain section of the flexible probe is as shown in formula (3): Among them A i is the coefficient matrix when the needle core extension is i.

2. The method for assisted puncture soft tissue viscoelasticity testing based on multi-curvature sample optimization according to claim 1, characterized in that: The puncture experiment described in step 1) is to use a pre-bent flexible probe to puncture soft tissue in a horizontal plane, and the pre-bent flexible probe is an actively controllable flexible probe in which a pre-bent needle core and a flexible needle sleeve are coaxially arranged; multiple curvature trajectories are obtained by changing the pre-bent state, which are recorded as multi-curvature characteristics; the pre-bent state is the relative elongation of the pre-bent needle core relative to the flexible needle sleeve, and changing the pre-bent state changes the contact form with the soft tissue.

3. The method for assisted puncture soft tissue viscoelasticity testing based on multi-curvature sample optimization according to claim 1, characterized in that: In step 3), the quasi-static equation based on the spring model is used to model the pre-bent flexible probe-soft tissue coupling model. The quasi-static model of the pre-bent flexible probe is established based on the interaction between the flexible needle and the soft tissue. The pre-bent needle core and the flexible needle sleeve are modeled as the needle body. The friction of the needle core is ignored. The flexible probe is defined as a cantilever beam. The probe is modeled using the Euler-Bernoulli theory, and a fourth-order ordinary differential equation is obtained as shown in formula (4). Where x is the feed amount of the flexible probe, ω is the deflection of the flexible probe, E is the Young's modulus of the flexible probe, I is the moment of inertia of the flexible probe, α is the angle between the bevel of the flexible needle tip and the soft tissue level, and F q is the tissue pressure on the needle tip bevel, F t / l is the vertical load on the needle bar, is the slope of the needle, is the bending moment of the needle, is the shear force of the needle; the solution process is to integrate formula (4) four times respectively, and obtain the value of the integral constant through the boundary conditions each time.

4. The method for assisted puncture soft tissue viscoelasticity testing based on multi-curvature sample optimization according to claim 1, characterized in that: The pseudo-spring coefficient described in step 4) is the value of the coefficient K obtained by solving the load on the needle rod based on the fourth-order ordinary differential equation; the pseudo-spring coefficient is the common coefficient of the influence of the soft tissue elastic coefficient and the viscosity coefficient on the deflection of the pre-bent flexible needle. In the auxiliary pre-bent flexible needle puncture soft tissue experiment, the deflection of the probe is affected by the elastic properties of the soft tissue, and the influence of the viscosity coefficient on the deflection is negligible. The pseudo-spring coefficient obtained is the comprehensive actual influence coefficient.

5. The method for assisted puncture soft tissue viscoelasticity testing based on multi-curvature sample optimization according to claim 1, characterized in that: The small batch gradient descent method described in step 4) is an optimization method for the coefficient K. The training set established is the deflection of the multi-curvature feature. The objective function established is the error between the needle bar deflection based on the pseudo-spring coefficient and the corresponding multi-curvature feature. Substituting it into the initial coefficient is K intial The pseudo-spring coefficient is obtained by fitting the single curvature characteristic through the fourth-order ordinary differential equation, and the objective function J(K) is as shown in formula (5); Where x n is the probe feed amount of the nth needle bar, y n is the x in the multi-curvature feature n The corresponding deflection value at ω K (x n ) is the parameter K and is substituted into the quasi-static model to solve the x n The corresponding deflection value at The small batch sample is to divide the data set into m sets, each set contains a curvature feature; The optimization method is to perform batch gradient descent on these sets one by one and update the parameter K; until the number of iterations reaches the threshold or the cost function drops to the minimum, the update process stops and the optimization coefficient K is obtained. final The coefficient K that minimizes the objective function J(K) is updated as shown in formula (6); Where (x p ,y p ) is the probe feed amount and corresponding deflection value contained in the multi-curvature feature in the pth set; η is the learning rate.

Images