Path planning method of puncture robot in lung free breathing state

Abstract

The invention discloses a path planning method for a puncture robot in a lung free breathing state, which comprises the following steps: firstly, analyzing the interaction force between a flexible needle and multi-layer soft tissues of a human body in different puncture stages, and establishing a kinematics model for interaction between the flexible needle and the multi-layer soft tissues by introducing a single-wheel vehicle model; then, a needle point position prediction model is built based on the nonlinear rod model, and the actual deflection posture of the flexible needle in the body is calculated in real time through force and torque parameters obtained by a force sensor at the tail end of the puncture mechanical arm. According to the method, a curve path generated by a flexible needle and a tissue interaction model is used for replacing a linear path in a traditional rapid search tree algorithm, and meanwhile, dynamic position changes of a target spot and an obstacle in a free breathing state are considered, so that dynamic obstacle avoidance is realized, and the puncture precision of a robot is improved. And finally, a lung puncture robot experiment platform in a free breathing state is constructed, and high-precision and high-efficiency puncture robot path planning can be realized in a free breathing environment.

Description

Technical Field

[0001] This invention relates to the field of medical puncture robot path planning technology. More specifically, it relates to a path planning method and system for a lung puncture robot under free breathing conditions, applicable to clinical procedures such as percutaneous puncture, biopsy, and local treatment. Background Technology

[0002] Cancer has become a major public health problem worldwide, with lung cancer ranking first in both incidence and mortality in my country. Lung biopsy is a commonly used clinical method in the treatment of lung cancer. However, this technique is characterized by its high dependence on the surgeon's experience, significant dynamic changes in the target location, and complex puncture path planning. Lung tissue undergoes periodic displacement under free breathing conditions, and the target lesion point will dynamically shift within a three-dimensional spatial range with the respiratory cycle. Without real-time compensation, puncture errors can easily occur, affecting the success rate and safety of the biopsy.

[0003] Existing path planning technologies mainly include static image-based planning methods and real-time navigation methods. Static image-based planning methods typically design paths based on preoperative CT images, making it difficult to adapt to dynamic changes caused by respiration. Meanwhile, traditional artificial potential field path planning methods are prone to getting trapped in local optima, while the fast search tree algorithm suffers from high randomness and long search time.

[0004] Furthermore, during lung puncture, the needle needs to traverse multiple layers of tissue, including the chest wall, fat layer, muscle tissue, and intercostal spaces. These tissues exhibit complex nonlinear mechanical behavior and multilayered structural characteristics, and the puncture path must avoid high-risk anatomical structures such as ribs, arteries, and veins. Existing methods typically lack comprehensive consideration of multilayered tissue structures, mechanical properties, and real-time respiratory cycles, resulting in insufficient reliability and safety in path planning.

[0005] Therefore, there is an urgent need for a complete path planning method and system that can simultaneously consider real-time respiratory dynamics, puncture needle posture constraints, obstacle avoidance of high-risk tissues, and path smoothness. Summary of the Invention

[0006] This invention aims to provide a path planning method and system for a lung puncture robot. By organically integrating an artificial potential field algorithm with a rapid exploring random tree (RRT) algorithm and combining it with a respiratory cycle dynamic model, it achieves safe, efficient, and real-time updatable puncture path planning, improving the accuracy and safety of lung puncture. The system establishes a multi-layered tissue model based on the complex thoracic cavity structure, calculates the target point displacement changes in real time using respiratory dynamics, and uses an artificial potential field (APF) to guide the search direction of the rapid exploring random tree (RRT) algorithm. This maintains the continuity of the puncture direction in high-risk, densely packed tissue areas, avoiding local extrema problems, and ultimately generates a safe, continuous, and real-time updatable path, enabling the robot to complete precise lung punctures even with free breathing.

[0007] The present invention adopts the following technical solution:

[0008] A path planning method and system for a lung puncture robot under free breathing conditions, including needle-tissue interaction modeling, APF-guided RRT path planning, respiratory motion model and robot vision system;

[0009] The needle-tissue interaction model module consists of a multi-layer tissue puncture mechanics model, a multi-layer tissue puncture kinematics model, and a needle tip position prediction model, which are used to describe the force and positional changes of the flexible needle in multi-layer tissue.

[0010] The multi-layer tissue puncture mechanical model is used to analyze the force situation during the actual puncture of multiple layers of human tissue (such as skin, fat, muscle, thoracic cavity and internal organs) by flexible needles.

[0011] Puncture kinematic models can infer the position of the needle tip when puncturing multi-layered tissues using sensor data, needle insertion speed, and needle rotation speed.

[0012] The needle tip position prediction model adopts the S. Awtar nonlinear rod model. Based on the nonlinear constitutive relation, the dynamic load-displacement transfer function of the needle tip can be established by solving the variational equation of the S. Awtar model, thereby obtaining the position information of the needle tip end.

[0013] The respiratory motion model module consists of a mechanical drive module, a pneumatic execution module, a biomimetic tissue construction, and a motion capture system. It simulates the characteristics of human respiratory motion through mechanical drive and pneumatic execution devices, converting respiratory frequency and tidal volume into mechanical reciprocating motion and gas volume changes, providing an experimental verification platform for dynamic path updates.

[0014] The mechanical drive module consists of a stepper motor and a linear module. The stepper motor drives the linear module via a coupling to achieve linear transmission.

[0015] Pneumatic actuator module. The pneumatic actuator module consists of a syringe, an air tube, and a hot water bag. It converts the human breathing frequency into the reciprocating motion frequency of the linear module and converts the tidal volume of human respiration into the inflation volume of the syringe into the hot water bag.

[0016] Bionic tissue. Bionic tissue consists of a gelatinous phantom and a bionic skin layer, used to simulate human anatomy.

[0017] The APF-guided RRT path planning module consists of RRT path planning and APF path planning.

[0018] During path search, attractive and repulsive potential fields are constructed using APF (Advanced Persistent Field Model). Gradient bias guides the RRT (Reverse Tracing and Restriction Time) sampling direction, and a dynamic step-size expansion strategy is combined to improve search efficiency and avoid local maxima. Based on the real-time displacement update of the respiratory motion model, adaptive path adjustment and dynamic obstacle avoidance are achieved.

[0019] The robot vision system consists of a puncture robotic arm, a six-dimensional force sensor, a puncture needle, and a vision system. The robotic arm and sensor enable the execution and feedback control of the puncture path, while the vision system captures the needle tip pose in real time to achieve path tracking and accuracy assessment.

[0020] The puncture robotic arm acts as the actuator, enabling the puncture needle to perform various motion controls, such as axial puncture, oblique puncture, and rotary needle insertion.

[0021] The six-dimensional force sensor is located between the end flange of the robotic arm and the puncture needle holder, and is used to record the mechanical data during the puncture process in real time.

[0022] During the puncture, the puncture needle body is allowed to deflect to a certain extent, and its movement trajectory can be arc-shaped. The needle tip can rotate 360° around its own axis to adapt to different puncture angles and path requirements.

[0023] The path capture system consists of a gelatin phantom, graph paper, a binocular camera, and a computer. When the flexible needle punctures the gelatin phantom on the experimental stage, the binocular camera captures the kinematic data of the flexible needle tip on the graph paper in real time, and the computer's active vision measurement algorithm fits the experimental data results.

[0024] This invention proposes a fast search tree path planning algorithm guided by the artificial potential field method based on the needle-multilayer tissue interaction model and respiratory motion model. It effectively solves the practical problems of low puncture accuracy and poor applicability caused by different tissue characteristics of the human body and free breathing during clinical lung puncture. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the force exerted during multi-layer tissue puncture.

[0026] Figure 2 This is a schematic diagram of a respiratory movement model.

[0027] Figure 3 This is a flowchart of the RRT path planning process guided by APF.

[0028] Figure 4 This is a schematic diagram of the path planning results (avoiding ribs and blood vessels).

[0029] Figure 5 A schematic diagram of the experimental platform for the robotic puncture system.

[0030] Figure 6 This is a schematic diagram illustrating the effect of the puncture robot path verification. Detailed Implementation

[0031] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide specific implementation methods and operating procedures; however, the scope of protection of the present invention is not limited to the following embodiments.

[0032] This invention provides a path planning method and system for a lung puncture robot under free breathing conditions. The system mainly consists of four modules: needle-tissue interaction modeling, APF-guided RRT path planning, respiratory motion model, and robotic arm vision system.

[0033] Module 1: Needle-Tissue Interaction Modeling

[0034] The needle-tissue interaction modeling module consists of three parts: a multi-layer tissue puncture mechanical model, a needle tip position prediction model, and a multi-layer tissue puncture kinematic model. It is used to accurately describe the mechanical response and needle tip pose changes during the puncture process in multi-layer biological tissues.

[0035] like Figure 1 The model shown illustrates the force characteristics during flexible needle puncture of multi-layered biological tissues (skin, fat, muscle, pleural cavity, etc.). The puncture process is divided into six stages, taking into account contact force, shear force, friction force, and resistance factors in sequence.

[0036] The puncture kinematic model establishes the needle tip pose estimation relationship based on the single-wheel vehicle model:

[0037] in, These are the coordinates of the tip of the needle. Let be the radius of the arc of the (i+1)th segment. The angle between the radius of the arc before entering the next tissue segment and the X-axis. The angle of incidence is denoted as .

[0038] Since biological tissues exhibit significant nonlinear constitutive properties, this system introduces S. Awtar cantilever beam theory to construct an enhanced mechanical model. By solving the variational equations of the S. Awtar model, the dynamic load-displacement transfer function of the needle tip can be established, thereby realizing the prediction of the needle tip pose.

[0039] Module Two: Respiratory Movement Model Module

[0040] Respiratory movement model such as Figure 2 As shown, this module is used to simulate the periodic motion changes of the lungs under free breathing conditions, providing experimental verification conditions for dynamic path updates.

[0041] The mechanical drive module consists of a 57-stepper motor and a high-precision linear module. The stepper motor drives the linear module via a coupling to achieve linear transmission, thereby moving the module slider. The stepper motor has a step angle of 1.8° and an output torque of 2.2 N·m. The high-precision linear module has a stroke of 200 mm and a repeatability of ±0.02 mm.

[0042] The pneumatic system consists of a syringe, an air delivery tube, and a hot water bag. A large-capacity syringe is directly connected to the slider of a linear module via a rigid connector, converting the slider's motion into the syringe piston's motion. The syringe outlet is sealed to the hot water bag via a silicone tube, with a one-way valve inside the tube to prevent backflow. The syringe piston's motion directly drives changes in the hot water bag's internal volume, simulating the inhalation and exhalation processes of the lungs.

[0043] A gelatin phantom measuring 50 mm × 50 mm × 200 mm was fixed to a radiator. A 3 mm diameter black glass sphere and a green cuboid were embedded within the phantom to simulate the target and obstacles in a puncture experiment. Finally, a 2 mm thick silicone bionic skin layer was applied over the phantom. The gelatin phantom was made by mixing gelatin particles and water in a 10:1 ratio, boiling the mixture, and then refrigerating it.

[0044] Module 3: RRT path planning module based on APF guidance

[0045] The entire path planning flowchart is as follows Figure 3 As shown, this module is the core part of the present invention, used to realize path planning based on the interaction modeling of flexible needles and multi-layer tissues in a free breathing environment.

[0046] First, target point and obstacle information are obtained from the DICOM image. Then, a traditional Path Response Theory (RRT) is constructed, and the interaction model of flexible needles and multi-layered tissues is integrated into the path planning. The improved RRT algorithm uses an arc-shaped path search method to obtain an initial reference path. The RRT path planning algorithm that integrates the interaction model of flexible needles and multi-layered tissues uses the arc path of the flexible needles to replace the straight path in the traditional RRT algorithm to connect the generated new nodes with the parent node, and ensures the continuous differentiability of the connection trajectory, so that the position and orientation of the connection points remain consistent.

[0047] An APF (Advanced Persistent Field) is established in this space, where the target point serves as the source of the attractive potential field, and obstacles such as bones and blood vessels serve as the source of the repulsive potential field. The potential field function can be expressed as:

[0048] in To attract the energy field, To repel the potential field, the path is controlled to move closer to the target and deviate in the obstacle avoidance direction, respectively.

[0049] The target region's location is updated in real-time based on the periodic changes in the target region's position provided by the respiratory motion model. Then, within each respiratory cycle, the path is dynamically adjusted according to the updated target region position to ensure the flexible needle avoids new obstacles and accurately reaches the target point. During this process, the repulsive force of obstacles also changes with respiratory motion; the algorithm updates the obstacle attraction in the artificial potential field to ensure the needle tip avoids other organs or tissues. On the new reference path node set, the point closest to each path node is selected as the new path node, and starting from the current position, a circular arc path is fitted through needle-tissue interaction modeling to connect to the next path node. Figure 4 This is a schematic diagram of the final path planning.

[0050] Module 4: Robotic Arm Execution and Path Tracking Module

[0051] like Figure 5 As shown, the robotic arm and vision system structure module consists of a puncture robotic arm, a six-dimensional force sensor, a puncture needle, and a vision system.

[0052] The collaborative robotic arm utilizes the Fao FR5 six-DOF collaborative robotic arm, which features high-precision position control and appropriate load capacity, with an end-effector repeatability of ±0.02 mm and a load capacity of 5 kg. The force sensor employs an ATI Mini45 six-dimensional force sensor, with a nonlinearity error of <0.5%FS and a repeatability error of <0.1%FS. The flexible puncture needle is a standard 22G biopsy needle with an outer diameter of 0.72 mm and a bevel angle of 15°. The vision system uses a Basler Ace 2 series binocular industrial camera with a sampling rate of 1 kHz, fixed to the side of the robotic arm via a tripod, approximately 1.5 m from the human model, with the lens optical axis at an angle of approximately 45° to the model surface. High-precision checkerboard coordinate calibration paper with a grid size of 1 mm × 1 mm is tightly adhered to the surface of the gelatinous phantom, and an acrylic pressure plate ensures a seamless fit between the grid and the phantom. The phantom was placed horizontally on a hot water bag of a respiratory motion model with a pre-set respiratory rate of 12 breaths / min, a cycle T of 5 seconds, and a target displacement amplitude of 15-20 mm. Then, using the control system of the Fao robotic arm, the puncture needle tip was aligned with the gelatin surface, with the puncture start point set at (50,0) and the end point at (50,150), tilted at 75°, and inserted at a speed of 10 mm / s to a depth of 100 mm. Simultaneously, a six-dimensional force sensor and a binocular camera were activated to record the mechanical and kinematic data in real time during the puncture process, used to verify the accuracy of path planning and prediction. Experimental results are as follows: Figure 6 As shown.

Claims

1. A method for path planning of a puncture robot under free breathing conditions, characterized in that, Includes the following modules: The needle-tissue interaction modeling module is used to establish the mechanical and kinematic relationship between flexible needles and multi-layered human tissues. This module includes: a multi-layered tissue puncture mechanical model to describe the force characteristics of the needle body when puncturing multiple layers of tissues such as skin, fat, muscle, thoracic cavity, lung tissue, bone, and blood vessels; a multi-layered tissue puncture kinematic model to calculate the position of the needle tip based on the needle insertion speed and needle rotation speed; and a needle tip position prediction model, based on the S. Awtar nonlinear rod model, which obtains the dynamic load displacement transfer function of the needle tip by solving variational equations, thereby predicting the pose of the needle tip. The respiratory motion model module is used to simulate the characteristics of human respiratory motion and provide real-time displacement change data of target points; this module includes a mechanical drive module, a pneumatic actuation module, a bionic tissue construction and motion capture system; The Rapid Search Tree (RRT) path planning module, guided by the Artificial Potential Field (APF) method, is used for dynamic path planning under free breathing conditions combined with a multi-layer tissue model. During the path search process, attractive and repulsive potential fields are constructed, and the RRT sampling direction is guided by gradient bias. Combined with a dynamic step size expansion strategy, path smoothness, obstacle avoidance, and adaptive updates are achieved. The robotic arm and vision system module are used to execute the puncture path and provide real-time feedback. The puncture robotic arm is linked with the vision system through a six-dimensional force sensor to monitor the force and posture of the puncture needle in real time, thereby achieving path tracking and accuracy assessment.

2. The puncture robot path planning method under free breathing conditions of the lungs according to claim 1, characterized in that, The needle tip position prediction model adopts the S. Awtar cantilever beam theory and obtains the bending response and end displacement of the flexible needle in multilayer tissue by establishing a nonlinear constitutive relationship.

3. The path planning method and system according to claim 1, characterized in that, The mechanical drive module in the breathing motion model module consists of a stepper motor and a linear module. The stepper motor drives the slider of the linear module through a coupling to achieve reciprocating linear transmission, so as to simulate the change of breathing frequency.

4. The puncture robot path planning method under free breathing conditions of the lungs according to claim 1, characterized in that, The pneumatic actuator module consists of a syringe, an air tube, and a hot water bag. The reciprocating motion of the syringe piston drives the change in gas volume in the hot water bag to simulate the inhalation and exhalation process.

5. The puncture robot path planning method under free breathing conditions of the lungs according to claim 1, characterized in that, In the APF-guided RRT path planning module, the attractive potential field With repulsive potential field These are used to control the direction in which the needle tip approaches the target and avoids obstacles, respectively, and their total potential field is: .

6. The puncture robot path planning method under free breathing conditions of the lungs according to claim 1, characterized in that, The path planning module combines the real-time displacement update of the respiratory motion model to achieve dynamic adjustment of the path node set and adaptive obstacle avoidance.

7. The puncture robot path planning method under free breathing conditions of the lungs according to claim 1, characterized in that, The puncture robotic arm is a six-degree-of-freedom collaborative robotic arm, with a six-dimensional force sensor and a puncture needle holder installed at its end, enabling axial puncture, oblique puncture, and rotary needle insertion motion control.

8. The puncture robot path planning method under free breathing conditions according to claim 1, characterized in that, The vision system is a binocular vision structure. It acquires three-dimensional motion data of the needle tip by collecting coordinate calibration paper images and combining them with an active vision measurement algorithm, which is used to verify the accuracy of path tracking.

9. The puncture robot path planning method under free breathing conditions of the lungs according to claim 1, characterized in that, The RRT path planning uses circular arc paths to connect and generate nodes, thereby maintaining the path's continuity and differentiability and ensuring the consistency of the needle tip's attitude.

10. The puncture robot path planning method under free breathing conditions according to claim 1, characterized in that, The system combines a flexible needle with a multi-layer tissue interaction model, a respiratory dynamics model, and an APF-guided RRT algorithm to achieve high-precision planning and real-time updating of the lung puncture path under free breathing conditions.

Images