Flexible puncture needle path planning method based on improved Bi-RRT algorithm

By improving the Bi-RRT algorithm, combining target bias and fast expansion strategies, the problems of randomness and inefficiency in flexible needle path planning are solved, and faster and safer path planning is achieved, which is suitable for real-time flexible puncture needle path planning requirements .

CN116784975BActive Publication Date: 2024-11-22ANHUI MEDICAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310461899.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-04-26
Publication Date
2024-11-22
Estimated Expiration
2043-04-26

AI Technical Summary

Technical Problem

The RRT algorithm has randomness and inefficiency in the existing flexible needle puncture path planning, which makes the path planning complex and difficult to quickly proceed, and is difficult to meet real-time needs.

Method used

Using a flexible puncture needle path planning method based on improved Bi-RRT algorithm, the target bias strategy and fast expansion strategy can reduce the calculation amount and path randomness, improve search speed and planning efficiency, including image recognition, environmental expansion, and target bias , pregrowth, rapid expansion, reaching the confluence point, deleting redundant nodes and smoothing processing, etc.

Benefits of technology

Faster convergence speed and higher efficiency path planning are achieved, reducing the number of inflection points, reducing the risk of puncture and improving safety.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116784975B_ABST
    Figure CN116784975B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of intelligent medical auxiliary robot control technology, in particular to a flexible puncture needle path planning method based on an improved Bi-RRT algorithm, the algorithm comprising the following steps: obtaining environmental information of a lesion site through a medical imaging device; establishing a puncture path model of a flexible needle; expanding the radius of obstacles in the environment to 1.5 times of the original radius to leave a safe distance; introducing a target bias strategy and finding a convergence point based on the target bias strategy; pre-growth to shorten the tree building distance; rapidly expanding from a starting point and a target point toward the convergence point, respectively, while avoiding obstacles; when the distance between the latest expanded node and the convergence point is less than a threshold, it is considered that the convergence point has been reached and the tree building is completed; removing redundant nodes on the path; and smoothing the new path. The present invention introduces a target bias strategy and a rapid expansion strategy to reduce the amount of calculation and complexity of calculation, accelerate the convergence speed, improve the efficiency of path planning, and obtain a planned path that can be applied to flexible needle puncture according to the constraints of the puncture path model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent medical auxiliary robot control, and in particular to a flexible puncture needle path planning method based on an improved Bi-RRT algorithm. Background Art

[0002] With the rapid development of minimally invasive surgery, puncture has been widely used in clinical practice as a minimally invasive surgery, which can complete tasks such as biopsy, brachytherapy, drug delivery and targeted therapy. Compared with traditional manual puncture, robot-assisted puncture has higher accuracy and stability, can avoid long-term direct exposure of doctors to radiation environment, can greatly reduce the labor intensity of doctors, and has received widespread attention and attention, and is the future development trend [1].

[0003] Currently, the puncture needles used in clinical practice are rigid needles. When the needle tip deviates from the target or there are nerves, blood vessels, important organs and other tissues on the puncture path, the doctor will manually adjust the needle tail to forcibly correct the deviation or avoid important tissues and organs, which will cause secondary harm to the patient, increase the patient's pain and increase the postoperative time. [2]

[0004] In order to solve the above problems, Webster et al. from Johns Hopkins University in the United States first proposed the concept of an asymmetric bevel-tipped flexible needle [3]. The flexible needle is made of nickel-titanium alloy. Its puncture path in the tissue can be one or more arcs, which can flexibly avoid nerves, blood vessels, important organs and other tissues, and accurately reach locations that traditional steel needles cannot reach. Due to the incomplete constraint of the needle tip [4][5], as well as the presence of sensitive tissues (such as nerves and blood) and obstacles (such as bones), the puncture trajectory of the flexible needle in the body requires high precision, and the precise control of the flexible needle requires high precision. Therefore, the flexible needle path planning is the basis for accurate puncture and is of great significance in the field of minimally invasive surgery.

[0005] Flexible needle path planning is the basis for precise control of medical robots. During robot-assisted puncture surgery, it is necessary to use imaging technology to plan a flexible needle insertion path according to the target. The principle is: in an environment with existing bone, blood vessels or organ obstacles, the path of the puncture needle should reasonably avoid obstacles, ensure that it does not collide with other organs of the patient while accurately reaching the target, and make the length of the puncture path as short as possible to relieve the patient's pain. Researchers have successively proposed probabilistic methods for flexible needle insertion path planning [6], objective function methods [7], artificial potential field methods [8] and inverse kinematics methods [9]. At present, reference

[10] combines RRT with reachability-guided sampling heuristics to propose a fast algorithm, which greatly increases the search efficiency. Reference [7] proposes an improved RRT algorithm based on environmental characteristics, combined with a path evaluation function including path length, number of arc segments, and safety, and sets the weight coefficient in the path evaluation function according to the obstacle situation in the environment. The corresponding optimal path is selected from the sample path through the evaluation function. Reference

[11] combined the old point tracking strategy and the extreme trend expansion strategy to propose a real-time flexible needle puncture path planning algorithm, which can dynamically adjust the path to achieve the real-time purpose. Reference

[12] based on the characteristics of flexible needle movement, * The algorithm is improved and the artificial potential field is applied to the obstacle avoidance strategy of the algorithm. However, due to the randomness of the traditional RRT algorithm, the determined path generated by the algorithm may encounter obstacles and the calculation convergence speed is slow, which is not suitable for real-time flexible puncture needle path planning.

[0006] References are as follows:

[0007] [1] Wang Guobiao, Peng Fangyu, Wang Shuxin, Zhu Shahong, Wang Tianmiao. Research progress of minimally invasive surgical robots - a review of the Shuangqing Forum on "Basic theories and key technologies of minimally invasive surgical robots and instruments"[J]. Chinese Science Foundation, 2009, 23(04): 209-214. DOI: 10.16262 / j.cnki.1000-8217.2009.04.006.

[0008] [2] Hu Hailong. Path planning of cannula flexible needle based on improved RRT algorithm[D]. Harbin University of Science and Technology, 2018.

[0009] [3]Webster III RJ, Kim JS, Cowan NJ, et al. Nonholonomic modeling of needle steering [J]. The International Journal of Robotics Research, 2006, 25(5-6): 509-525.

[0010] [4]Webster RJ,Memisevic J,Okamura A M.Design considerations for robotic needle steering[C] / / Proceedings of the 2005IEEE International Conference on Robotics and Automation.IEEE,2005:3588-3594.

[0011] [5]Webster III RJ, Kim JS, Cowan NJ, et al. Nonholonomic modeling of needle steering[J]. The International Journal of Robotics Research, 2006, 25(5-6):509-525.

[0012] [6]Alterovitz R,Patil S,Derbakova A. Rapidly-exploring roadmaps: Weighing exploration vs.refinement in optimal motion planning[C] / / 2011IEEEInternational Conference on Robotics and Automation.IEEE,2011:3706-3712.

[0013] [7] Li Xia. Research on robot-assisted flexible needle soft tissue puncture system[D]. Xiangtan University, 2016.

[0014] [8] Liu Xiaoyan. Acupuncture soft tissue trajectory planning and experimental research based on potential energy field[D]. Tianjin University, 2010.

[0015] [9] Adagolodjo Y, Goffin L, De Mathelin M, et al. Robotic insertion of flexible needle in deformable structures using inverse finite-elementsimulation[J]. IEEE Transactions on Robotics, 2019, 35(3): 697-708.

[0016]

[10] Patil S,Alterovitz R.Interactive motion planning for steerableneedles in 3D environments with obstacles[C] / / 20103rd IEEE RAS&EMBSinternational conference on biomedical robotics and biomechatronics.IEEE,2010:893-899.

[0017]

[11] Wang Yuanzheng. Dynamic puncture path planning of puncture probe based on ultrasound image guidance[D]. Harbin Institute of Technology, 2020. DOI: 10.27061 / d.cnki.ghgdu.2020.002940.

[0018]

[12] Zhang Y, Qi Z, Zhang H. An Improved RRT Algorithm Combining MotionConstraint and Artificial Potential Field for Robot-Assisted Flexible NeedleInsertion in 3D Environment [C] / / 20213rd International Conference on Industrial Artificial Intelligence (Iai). IEEE, 2021: 1-6. Summary of the invention

[0019] In view of the problems that the RRT algorithm in the existing flexible needle puncture path planning is random and inefficient, the planned path is complex, and it is difficult to quickly plan the path, the present invention proposes a flexible puncture needle path planning method based on an improved Bi-RRT algorithm. Based on the target bias strategy and the rapid expansion strategy, the calculation amount and path randomness are reduced, while the search speed is accelerated to improve the efficiency of path planning.

[0020] The technical problem to be solved by the present invention is achieved by adopting the following technical solutions:

[0021] The flexible puncture needle path planning method based on the improved Bi-RRT algorithm includes the following steps:

[0022] Step (S1) Image recognition:

[0023] Obtain environmental information of the lesion site through medical imaging equipment to determine information such as the location of the needle point, target point, and obstacles;

[0024] Step (S2) establishing a puncture model:

[0025] Establish a puncture path model for the flexible needle;

[0026] Step (S3) Environment Expansion:

[0027] Perform environmental expansion, increasing the obstacle radius by 1.5 times on the same map;

[0028] Step (S4) target offset:

[0029] The same step size and threshold as those in the traditional RRT algorithm are set, a target bias strategy is introduced, and the convergence point is found based on the target bias strategy;

[0030] Step (S5) pre-growth:

[0031] Pre-growth shortens the distance to establish a tree;

[0032] Step (S6) Rapid expansion:

[0033] Rapidly expand from the starting point and the target point towards the meeting point while avoiding obstacles;

[0034] Step (S7) reaches the junction:

[0035] The bidirectional expansion tree stops building when it reaches the confluence point or the distance to the confluence point is less than the threshold, forming a collision-free path from the initial point to the target point;

[0036] Step (S8) deletes redundant nodes:

[0037] Remove redundant nodes on the path;

[0038] Step (S9) Smoothing Processing:

[0039] Smooth the new path and finally get a smooth curve;

[0040] Step (S10) determines whether the fitted path meets the puncture constraint condition. If so, the current path is output. If not, the path planning is terminated or the step size or threshold in the algorithm is adjusted, and steps (S5) to (S9) are continued.

[0041] Preferably, the puncture path model of the flexible needle in step (S2) is expressed as:

[0042] Path=[(X1, D1, l1, r1), (X2, D2, l2, r2), (X n , D n , l n , r n )]

[0043] The path Path is composed of n segments, where X i represents the starting point of the i-th path, D i Indicates the direction of the tangent line between the i-th arc and the previous arc, l i represents the length of the i-th arc, r i Represents the radius of the i-th arc.

[0044] Preferably, in step (S4), a sampling convergence point P is selected between the starting point and the target point. converge , as the expansion direction of the two trees; when the starting point O init (x init ,y init ) and the target point O goal (x goal ,y goal ) When no obstacle is detected at the midpoint of the connecting line, the midpoint is directly selected. as a meeting point;

[0045] When an obstacle is detected at the midpoint of the line connecting the starting point and the target point, take the midpoint as the center of the circle and 1 / 50 of the length of the line as the radius R. Draw a perpendicular bisector of the line through the center point. If the intersection of the perpendicular bisector and the circle is free from obstacles, take that point as P. converge If the intersection point cannot get rid of the obstacle, the radius will be doubled until it gets rid of the obstacle.

[0046] Preferably, the midpoint of P is on the obstacle. converge The coordinates are calculated as follows:

[0047]

[0048] Solve the system of equations and take the point not in the obstacle as P converge .

[0049] Preferably, step (S5) specifically includes:

[0050] Start from the starting point and the target point respectively, and expand along the line connecting the two points with a step length. If no collision occurs, continue to expand outward with a step length until a collision occurs, then return to the previous node.

[0051] Preferably, step (S6) specifically includes:

[0052] After pre-growth, the latest node of the two trees is the new starting point P init With the target point P goal , at the starting point P init and P converge Continue to expand the connection line at a multiple of the step length. If you reach P i If a collision occurs at point Pi-1 As the center, draw a circle with twice the step length as the radius r, collect 8 points on the circle at 45° intervals, perform collision detection on these 8 points respectively, and compare the collected points with P converge distance, select and P i-1 The line has no collision with obstacles and the distance is P converge The nearest point is P new , which is the latest expanded node; at the same time, from P goal Departures are also expanded in the same way as above.

[0053] Preferably, step (S7) specifically includes: connecting P new With P converge ,If there is an obstacle on the connecting line, repeat step (S7) until the distance is less than the threshold, and the search ends.

[0054] Preferably, step (S8) specifically includes:

[0055] Create redundant points on the path: init As the first head node P s , connect P s and its subsequent node P s+2 If there is no obstacle between the two points, then connect P s+2 The next node P s+3 , traverse in sequence, if a collision occurs, then P s+2 The previous node P s+1 As the new head node, continue to connect to the next node to detect collision, and repeat this step until you reach P converge , connect all head nodes and target nodes to form a new path.

[0056] Preferably, step (S9) specifically includes:

[0057] First, the first arc is drawn with the first two nodes as the two ends of the arc, and the radius is set to R1=50. The opening direction is consistent with the curve fitted by the first three nodes obtained in step (S8). The following arcs must meet the condition of being tangent to the previous arc, and the endpoints must be nodes on the tree.

[0058] Preferably, in step (S10), judging whether the current path meets the constraint conditions of the puncture path model includes:

[0059] (1) The sum of each path segment is less than or equal to the total length of the puncture needle; (2) The arcs must be tangent to each other, and the opening directions of the connected arcs must be inconsistent; (3) The curvature of each arc segment is less than or equal to the maximum bending curvature required by the stiffness of the flexible needle.

[0060] The beneficial effects of the present invention are:

[0061] The present invention provides a flexible puncture needle path planning method based on an improved Bi-RRT algorithm, which adds a target bias strategy and a rapid expansion strategy to the traditional Bi-RRT algorithm, has a faster convergence speed and higher efficiency; post-processes the path to reduce the number of inflection points, reduce the risk of puncture, and has higher safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The present invention is further described below in conjunction with the accompanying drawings and embodiments:

[0063] Figure 1 is a flow chart of the method of the present invention;

[0064] Figure 2 is a flow chart of step (S4) of the present invention;

[0065] Figure 3 is a flow chart of steps (S5), (S6), and (S7) of the present invention;

[0066] Figure 4 This is a schematic diagram of the algorithm search in step (S5) of the present invention;

[0067] Figure 5 It is a flow chart of step (S8) of the present invention. DETAILED DESCRIPTION

[0068] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further described below with reference to the accompanying drawings and embodiments.

[0069] like Figure 1 As shown, the flexible puncture needle path planning method based on the improved Bi-RRT algorithm includes the following steps:

[0070] Step (S1) Image recognition.

[0071] Medical imaging equipment such as CT, MRI, and ultrasound are used to obtain environmental information about the lesion site, thereby determining information such as the location of needle points, target points, and obstacles, and placing them in the coordinate plane through modeling.

[0072] Step (S2) establishes a puncture model.

[0073] Establish a puncture path model for a flexible needle.

[0074] The path geometry of the flexible needle during the puncture process is a single arc or multiple arcs or a straight line path: when the starting point O init With the target point O goalWhen there is no obstacle between the starting point and the target point, the obtained path is a straight line path from the starting point to the target point; when there is an obstacle between the starting point and the target point, the improved Bi-RRT algorithm is used to find points that can avoid the obstacle, and the points are connected to finally obtain a multi-segment composite path. The obtained path also needs to trim redundant points and smoothen to finally obtain a feasible path.

[0075] In practice, there are fewer straight paths, and more are combinations of multiple composite paths. The path model must meet the deformation constraints of the flexible needle: a. The sum of each path segment is less than or equal to the total length of the puncture needle; b. The arcs must be tangent, and the opening directions of the connected arcs are inconsistent; c. The curvature of each arc segment is less than or equal to the maximum bending curvature required by the stiffness of the flexible needle. The puncture path model Path can be described as:

[0076] Path=[(X1, D1, l1, r1), (X2, D2, l2, r2), (X n , D n , l n , r n )]

[0077] The path Path is composed of n segments, where X i represents the starting point of the i-th path, D i Indicates the direction of the tangent line between the i-th arc and the previous arc, l i represents the length of the i-th arc, r i Represents the radius of the i-th arc.

[0078] Step (S3) the environment expands.

[0079] Perform environmental expansion, increasing the obstacle radius by up to 1.5 times on the same map.

[0080] In practical applications, we should leave a safe distance between the flexible needle and the obstacle to deal with some unexpected events, such as deviation between the obstacle and the needle tip, and to ensure that there is no collision after arc fitting. Before searching the path, the obstacle area is expanded to a corresponding degree during the simulation process, and the obstacle radius is expanded to 1.5 times the original.

[0081] Step (S4) target offset. The specific process is as follows Figure 2 shown.

[0082] The same step size and threshold as those in the traditional RRT algorithm are set, a target bias strategy is introduced, and the convergence point is found based on the target bias strategy.

[0083] The step size l is usually set to one tenth of the edge length of the environment map, and the threshold d (nodes close to this threshold are considered almost the same) is set to one half of the step size l.

[0084] In view of the undirectedness of the traditional RRT algorithm search, a target bias strategy is introduced to select a sampling convergence point P between the starting point and the target point. converge , as the expansion direction of the two trees, so that the expansion of the algorithm is more directional and the amount of calculation is reduced.

[0085] When the starting point O init (x init ,y init ) and the target point O goal (x goal ,y goal ) When no obstacle is detected at the midpoint of the connecting line, the midpoint is directly selected. As the meeting point; when an obstacle is detected at the midpoint of the line connecting the starting point and the target point, take the midpoint as the center of the circle, and 1 / 50 of the length of the line as the radius R, and make a perpendicular bisector of the line through the center point. If the intersection of the perpendicular bisector and the circle has been clear of obstacles, then this point is taken as P converge If the intersection point cannot get rid of the obstacle, the radius will be doubled until it gets rid of the obstacle.

[0086] Take that point as P converge . The equations of the line and the circle are as follows:

[0087]

[0088] Find the equations that can be finally solved to get P converge Coordinates, the final two solutions are selected according to the actual situation: if one solution is within the obstacle, it is discarded; if neither is within the obstacle, any point is selected as P converge .

[0089] The pseudo code of this process is as follows:

[0090]

[0091] Step (S5) pre-growth. The specific process is as follows Figure 3 as well as Figure 4 shown.

[0092] Pre-growth shortens the tree building distance.

[0093] Start from the starting point and the target point respectively, and expand along the line connecting the two points with a step length. If no collision occurs, continue to expand outward with a step length until a collision occurs, then return to the previous node. This can shorten the search distance, save time, and reduce the possibility of collision in subsequent growth.

[0094] Step (S6) is rapid expansion. The specific process is as follows Figure 3 shown.

[0095] Rapidly expand from the starting point and the target point towards the meeting point while avoiding obstacles.

[0096] After pre-growth, the latest node of the two trees is the new starting point P init With the target point P goal , at the starting point P init and P converge Continue to expand the connection line at a multiple of the step length. If you reach P i If a collision occurs at point P i-1 As the center, draw a circle with twice the step length as the radius r, collect 8 points on the circle at 45° intervals, perform collision detection on these 8 points respectively, and compare the collected points with P converge distance, select and P i-1 The line has no collision with obstacles and the distance is P converge The nearest point is P new , which is the latest expanded node; at the same time, from P goal Departures are also expanded in the same way as above.

[0097] Step (S7) reaches the junction. The specific process is as follows Figure 3 shown.

[0098] The bidirectional expansion tree stops building when it reaches the confluence point or the distance to the confluence point is less than a threshold, forming a collision-free path from the initial point to the target point.

[0099] Connect P new With P converge ,If there is an obstacle on the connecting line, repeat step (S7) until the distance is less than the threshold, and the search ends.

[0100] The pseudo codes of steps (S6) and (S7) are as follows:

[0101]

[0102]

[0103] Step (S8) deletes redundant nodes. The specific process is as follows Figure 5 shown.

[0104] Remove redundant nodes on the path: init As the first head node P s , connect P s and its subsequent node P s+2 If there is no obstacle between the two points, then connect P s+2 The next node P s+3 , traverse in sequence, if a collision occurs, then P s+2 The previous node P s+1As the new head node, continue to connect to the next node to detect collision, and repeat this step until you reach P converge , connect all head nodes and target nodes to form a new path.

[0105] The pseudo code of this process is as follows:

[0106]

[0107]

[0108] Step (S9) Smoothing process.

[0109] The new path is smoothed to obtain a smooth curve.

[0110] First, draw the first arc with the first two nodes as the two ends of the arc, set the radius to R1=50, and the opening direction is consistent with the curve fitted by the first three nodes obtained before S8. The following arcs must meet the condition of being tangent to the previous arc, and the endpoints must be nodes on the tree.

[0111] Step (S10) determines whether the fitted path meets the puncture constraint condition. If so, the current path is output. If not, the path planning is terminated or the step size or threshold in the algorithm is adjusted, and steps (S5) to (S9) are continued.

[0112] The constraints for determining whether the current path meets the puncture path model include: (1) the sum of each path segment is less than or equal to the total length of the puncture needle; (2) the arcs must be tangent to each other, and the opening directions of the connected arcs are inconsistent; (3) the curvature of each arc segment is less than or equal to the maximum bending curvature required by the stiffness of the flexible needle.

[0113] The present invention improves the traditional Bi-RRT algorithm: first, a target bias strategy is added to make the expansion of the random tree more directional, thereby reducing the path-finding time; second, a fast expansion strategy is added to improve the expansion efficiency while preventing the algorithm from falling into the local optimum; the obtained points are screened to reduce redundancy. The obtained nodes can be used for path planning of the flexible needle puncture robot after smooth connection.

[0114] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A flexible puncture needle path planning method based on an improved Bi-RRT algorithm, characterized in that: The following steps are involved: Step S1: Image recognition: Obtain environmental information of the lesion site through medical imaging equipment to determine information such as the location of the needle point, target point, and obstacles; Step S2: Establishing the puncture model: Establish a puncture path model for the flexible needle; Step S3 environment expansion: Perform environmental expansion, increasing the obstacle radius by 1.5 times on the same map; Step S4 target offset: The same step size and threshold as those in the traditional RRT algorithm are set, a target bias strategy is introduced, and the convergence point is found based on the target bias strategy; Step S5: Pre-growth: Pre-growth shortens the distance to establish a tree; Step S6: Quick expansion: Rapidly expand from the starting point and the target point towards the meeting point while avoiding obstacles; Step S7: reaching the meeting point: The bidirectional expansion tree stops building when it reaches the confluence point or the distance to the confluence point is less than the threshold, forming a collision-free path from the initial point to the target point; Step S8: Delete redundant nodes: Remove redundant nodes on the path; Step S9: Smoothing process: Smooth the new path and finally get a smooth curve; Step S10 determines whether the fitted path meets the puncture constraint condition. If so, the current path is output. If not, the path planning is terminated or the step size or threshold in the algorithm is adjusted, and steps S5 to S9 are continued. The puncture path model of the flexible needle in step S2 is expressed as: The path Path is composed of n segments, where X i represents the starting point of the i-th path, D i Indicates the direction of the tangent line between the i-th arc and the previous arc. l i represents the length of the i-th arc, r i represents the radius of the i-th arc; The path geometry of the flexible needle during the puncture process is a single arc or multiple arcs. When there is an obstacle between the starting point and the target point, the improved Bi-RRT algorithm is used to find points that can avoid the obstacle, and the points are connected to finally obtain a multi-segment composite path. The obtained path also needs to trim redundant points and smoothen to finally obtain a feasible path. In step S10, it is determined whether the current path meets the constraint conditions of the puncture path model, including that the arcs must be tangent to each other and the opening directions of the connected arcs are inconsistent.

2. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 1, characterized in that: In step S4, a sampling convergence point P is selected between the starting point and the target point. converge , as the expansion direction of the two trees; when the starting point O init (x init ,y init ) and the target point O goal (x goal ,y goal ) When no obstacle is detected at the midpoint of the line, directly take the midpoint ( , ) as a meeting point; When an obstacle is detected at the midpoint of the line connecting the starting point and the target point, take the midpoint as the center of the circle and 1 / 50 of the length of the line as the radius R. Draw a perpendicular bisector of the line through the center point. If the intersection of the perpendicular bisector and the circle is free from obstacles, take that point as P. converge If the intersection point cannot get rid of the obstacle, the radius will be doubled until it gets rid of the obstacle.

3. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 2, characterized in that: P with its midpoint on the obstacle converge The coordinates are calculated as follows: Solve the system of equations and take the point not in the obstacle as P converge .

4. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 1, characterized in that: Step S5 specifically includes: Start from the starting point and the target point respectively, and expand along the line connecting the two points with a step length. If no collision occurs, continue to expand outward with a step length until a collision occurs, then return to the previous node.

5. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 1, characterized in that: Step S6 specifically includes: After pre-growth, the latest node of the two trees is the new starting point P init With the target point P goal , at the starting point P init and P converge Continue to expand the connection line at a multiple of the step length. If you reach P i If a collision occurs at point P i-1 As the center, draw a circle with twice the step length as the radius r, collect 8 points on the circle at 45° intervals, perform collision detection on these 8 points respectively, and compare the collected points with P converge distance, select and P i-1 The line has no collision with obstacles and the distance is P converge The nearest point is P new , which is the latest expanded node; at the same time, from P goal Departures are also expanded in the same way as above.

6. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 1, characterized in that: Step S7 specifically includes: connecting P new With P converge If there is an obstacle on the line, repeat step S7 until the distance is less than the threshold and the search ends.

7. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 1, characterized in that: Step S8 specifically includes: Create redundant points on the path: init As the first head node P s , connect P s and its subsequent node P s+2 If there is no obstacle between the two points, then connect P s+2 The next node P s+3 , traverse in sequence, if a collision occurs, then P s+2 The previous node P s+1 As the new head node, continue to connect to the next node to detect collision, and repeat this step until you reach P converge , connect all head nodes and target nodes to form a new path.

8. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 1, characterized in that: Step S9 specifically includes: First, draw the first arc with the first two nodes as the two ends of the arc, set the radius to R1=50, and the opening direction is consistent with the curve fitted by the first three nodes obtained in step S8. The following arcs must meet the condition of being tangent to the previous arc, and the endpoints must be nodes on the tree.

9. The flexible puncture needle path planning method based on the improved Bi-RRT algorithm according to claim 1, characterized in that: In step S10, judging whether the current path meets the constraint conditions of the puncture path model also includes: (1) The sum of each path segment is less than or equal to the total length of the puncture needle; (2) The curvature of each arc segment is less than or equal to the maximum bending curvature required by the stiffness of the flexible needle.