Hemostatic clamping control method and system for cardiovascular minimally invasive surgery

Abstract

The invention discloses a hemostasis clamping control method and system for a cardiovascular minimally invasive surgery, and relates to the technical field of cardiovascular minimally invasive surgeries, and the method comprises the steps: 1, collecting an image through a 4k endoscope, extracting a bleeding point region through a U-Net + + segmentation model, and recognizing a bleeding type through the combination of HOG features and an SVM classifier; 2, setting an initial opening degree through a displacement sensor, and adjusting a clamping angle by referring to the diameter of the blood vessel; thirdly, the electric clamping tool is started, the actual clamping force is fed back through a pressure sensor, and the rotating speed of a motor is dynamically adjusted to control the clamping force error; fourthly, the blood flow velocity is calculated through an optical flow method, tissue oxygen supply is monitored by combining a blood oxygen saturation sensor and a blood oxygen saturation sensor, and whether hemostasis succeeds or not is judged; 5, switching to a constant force maintaining mode after hemostasis succeeds, and reducing the clamping force according to the speed gradient after the operation is finished. The hemostasis control precision and stability are improved, the safety protection mechanism is perfected, the adaptability and practicability of the system are enhanced, and the clinical value is remarkable.

Description

Technical Field

[0001] This invention relates to the field of cardiovascular minimally invasive surgery technology, and in particular to a hemostatic clamping control method and system for cardiovascular minimally invasive surgery. Background Technology

[0002] Cardiovascular minimally invasive surgery has become the mainstream treatment for coronary heart disease and peripheral vascular disease due to its advantages such as minimal trauma and rapid recovery. However, bleeding control during surgery has always been a clinical challenge. The surgical area is small, the blood vessels are mostly between 1-10 mm in diameter, and they are surrounded by vital organs and nerve tissue, which significantly limits traditional hemostasis methods. Currently, the commonly used manual forceps hemostasis relies on the surgeon's experience, and the clamping force is entirely controlled by feel. Too little force can easily lead to hemostasis failure, while too much force may damage the vascular endothelium, causing complications such as thrombosis or vascular rupture. Especially during long surgeries, surgeon hand fatigue can increase the risk of operational errors.

[0003] While existing automated hemostasis equipment can partially replace manual operations, its technological maturity is insufficient. Most devices can only perform clamping actions according to preset parameters, lacking the ability to adapt to changes in the surgical environment. Fluctuations in temperature and air pressure can lead to a decrease in clamping force accuracy, and fatigue deformation of mechanical components after prolonged use can also affect control stability. In terms of bleeding point identification, traditional equipment mostly relies on a single endoscopic image, which is affected by blood reflection and tissue obstruction, making it difficult to accurately locate the bleeding point and determine the bleeding type, resulting in unreasonable selection of clamping angle and force. At the same time, existing systems lack a comprehensive damage prevention mechanism and cannot dynamically adjust safety thresholds based on parameters such as vessel diameter and compressive strength. Furthermore, hemostasis effect monitoring relies solely on a single blood flow indicator, easily overlooking issues such as insufficient tissue oxygen supply, affecting surgical safety.

[0004] Furthermore, the system integration and reliability of existing equipment need improvement. The clamping actuators are mostly custom-designed, unable to adapt to the needs of different surgical sites, and changing the clamp head is time-consuming. The control unit lacks data encryption and remote collaboration functions, making it difficult to meet medical data security standards and clinical teaching needs. Emergency fault handling capabilities are weak; in the event of sudden situations such as tool slippage or motor jamming, manual intervention is often relied upon, and the delayed response can easily delay hemostasis. These technical deficiencies result in a low adoption rate of existing equipment in clinical applications, making it difficult to meet the high-precision, high-safety, and highly adaptable hemostasis control requirements of minimally invasive cardiovascular surgery. Summary of the Invention

[0005] The present invention proposes a hemostasis clamping control method and system for cardiovascular minimally invasive surgery to solve the problems mentioned in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a hemostatic clamping control method for minimally invasive cardiovascular surgery, comprising the following steps: Step 1: Bleeding Point Localization and Feature Recognition Steps: Surgical field images are acquired using a 4K endoscope. The acquired images are processed using the U-Net++ segmentation model to extract the bleeding point region. The bleeding type is identified by combining HOG features and an SVM classifier. At the same time, the distance from the bleeding point to the clamping tool is measured using a laser rangefinder. Step 2: Clamping parameter initialization steps: Preset the clamping force baseline value according to the identified bleeding type; set the initial opening degree of the clamping tool through the displacement sensor; adjust the clamping angle based on the blood vessel diameter of the surgical site obtained by measuring through endoscopic images; Step 3: Adaptive clamping control steps: Start the electric clamping tool, which is driven by a stepper motor; control the clamping tool to approach the bleeding point according to the preset parameters, and provide real-time feedback of the actual clamping force through the pressure sensor installed on the clamping tool; compare the actual clamping force with the preset reference value, and dynamically adjust the motor speed according to the comparison result; Step 4: Monitoring the hemostasis effect: Analyze the blood flow status of the bleeding point through endoscopic images, specifically using optical flow to calculate blood flow velocity; and monitor the oxygen supply to the local tissue using a blood oxygen saturation sensor. Step 5: Clamping State Maintenance and Release Steps: After successful hemostasis, the control device switches to constant force maintenance mode and compensates for the mechanical drift of the clamping tool through the PID controller; after the operation is completed, the clamping force is reduced according to the preset rate gradient until the clamping tool is completely released.

[0007] Furthermore, it also includes a dynamic clamping force calibration step, which is integrated throughout the entire process of adaptive clamping control and constant force maintenance. This step improves control accuracy by collaboratively correcting clamping force deviations through multiple parameters. First, surgical environment parameters are collected in real time using temperature and pressure sensors, while simultaneously recording the continuous working time of the clamping tool. The calibrated clamping force calculation formula is as follows: ,in, To calibrate the clamping force, To preset the clamping force reference value, Temperature coefficient; This represents the difference between the actual surgical environment temperature and the standard temperature. This is the difference between the actual air pressure and the standard air pressure; the motor output torque is adjusted via a PWM signal after each calibration.

[0008] Furthermore, it also includes a damage prevention protection step, which is initiated after the clamping parameters are initialized. This step achieves dual protection for both blood vessels and tissues through layered calculation of the safety threshold. First, the diameter of the blood vessel at the surgical site is measured using endoscopic images to calculate the clamping contact area. Then, the maximum compressive strength is determined based on the blood vessel type. Finally, the safe clamping force threshold is calculated using a formula. ,in, For the safety clamping force threshold, A represents the maximum compressive strength of the blood vessel at the surgical site, and A is the contact area between the clamping tool and the blood vessel. For the clamping angle, D is the blood vessel diameter correction factor; D is the blood vessel diameter; a three-level early warning mechanism is set: when the actual clamping force reaches... When the clamping force reaches 70%, the control panel displays a yellow warning; when it reaches 80%, the clamping force growth rate is reduced to 0.05 N / s; when it reaches 90%, a red warning is issued and the clamping force is locked.

[0009] Furthermore, in step 1, multimodal fusion and 3D reconstruction techniques are used to improve the accuracy of bleeding point localization. First, 4K endoscopic images and near-infrared imaging data are acquired simultaneously. Narrow-band filters are used to filter stray light and enhance the grayscale difference between hemoglobin and surrounding tissues. The two images are then input into an image fusion module, which uses a Laplacian pyramid fusion algorithm. The bottom layer retains the detailed information of the endoscopic image, while the top layer overlays the bleeding point features of the near-infrared image. The fused stereo image pairs are acquired through a binocular vision system, the disparity map is calculated, and the 3D coordinates of the bleeding point are reconstructed. Combined with the robotic arm coordinate system of the surgical robot, which uses an absolute encoder, the coordinates of the bleeding point are converted into target points in the robotic arm base coordinate system to generate the motion path of the gripping tool.

[0010] Furthermore, in step 3, the adaptive clamping control employs a composite control algorithm combining fuzzy PID and feedforward compensation; firstly, the clamping force deviation is defined. and the rate of change of deviation ,in To determine the actual clamping force, the fuzzy sets of both are divided into seven levels: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large. The membership function adopts a triangular distribution. A fuzzy rule base containing 25 rules is established, and a feedforward compensation term is introduced. Based on the distance change rate measured by the laser rangefinder, the pre-torque required by the motor is calculated in advance to compensate for the lag caused by motion inertia.

[0011] Furthermore, step 4, in monitoring hemostasis effectiveness, employs multi-indicator fusion and dynamic judgment logic; four core monitoring indicators are set: first, blood flow velocity, calculated using optical flow to measure pixel movement velocity in the bleeding area; second, bleeding area, segmented in real-time using a U-Net++ model and calculated as the area reduction ratio; third, local oxygen saturation, monitored by a fiber optic oxygen probe to measure surrounding tissues at the clamping site; and fourth, tissue deformation at the clamping site, analyzed using endoscopic images to assess the degree of vessel wall indentation. A dynamic judgment rule is established: in the initial stage of surgery, if the blood flow velocity is sufficient... Preliminary hemostasis is determined when the clamping force is ≤0.5cm / s and the bleeding area is reduced by ≥90%. During the stable phase, the blood oxygen saturation must be ≥92% and the tissue deformation ≤0.2mm; otherwise, it is considered excessive clamping. During the sustained phase, at least two of the four indicators must be met simultaneously and the results must be sustained for more than 3 seconds to determine successful hemostasis. If the criteria are not met, the system will automatically initiate an adjustment procedure: first, increase the clamping force by 0.2N, maintain it for 1 second, and then check the indicators. If the criteria are still not met, adjust the clamping angle by 5° and increase the clamping force by 0.2N again, up to a maximum of 3 adjustments. If the adjustment still fails, a manual intervention prompt will be triggered.

[0012] Furthermore, it includes emergency intervention and fault self-repair steps, defining three types of emergency events: tool slippage, secondary bleeding, and motor failure. In the event of an emergency, a three-level response is immediately initiated: Level 1 braking, where the motor stops moving within 0.1 seconds and maintains the current clamping state; Level 2 alarm, where a buzzer sounds continuously at a 2kHz frequency, a red indicator light on the surgical console flashes, and the fault type and treatment suggestions are displayed on the touchscreen; Level 3 self-repair, where for tool slippage, the system automatically increases the clamping force slightly by 0.3-0.5N and fine-tunes the angle by 3° to attempt re-clamping; for slight motor jamming, a reverse pulse signal is output for 0.05 seconds to release the jamming. If the repair fails, the motor is locked and awaits manual intervention. Furthermore, including: The image acquisition and recognition unit consists of a 4K endoscope, a near-infrared imaging module, a binocular vision component, and an image processor. The 4K endoscope uses a CMOS image sensor. The near-infrared imaging module contains a 700-900nm LED light source. The left and right cameras of the binocular vision component are triggered synchronously. The image processor uses the NVIDIA Jetson AGX Orin platform. The clamping execution unit consists of a modular electric clamping clamp, a stepper motor drive module, and a displacement detection component. The sensing feedback unit integrates a pressure sensor, a blood oxygen saturation sensor, a temperature and pressure sensor, and a current sensor; the blood oxygen saturation sensor is a fiber optic probe type, measured by spectroscopy; the temperature and pressure sensor uses an integrated chip to simultaneously output temperature and pressure data to monitor the motor's operating status. The control unit is responsible for image preprocessing, sensor signal acquisition, and motor pulse output; the control unit supports real-time data caching and long-term storage. The human-computer interaction unit displays surgical field images, clamping force curves, sensor parameters, and system status, and supports doctors to manually set parameters such as clamping force and opening degree; the foot switch has three positions: clamping, releasing, and fine adjustment; the audible and visual alarm device includes a buzzer and red, yellow, and green indicator lights, with the buzzer frequency at 2kHz and the three-color lights corresponding to emergency fault, warning, and normal status, respectively.

[0013] Furthermore, the clamping execution unit adopts a redundant design and intelligent maintenance mechanism; the motor drive module has a built-in dual-path drive circuit, and when the main circuit fails, the backup circuit automatically switches within 0.2 seconds; an automatic grease replenishment device is installed inside the clamp body to maintain the flexibility of the transmission components; a clamp head life management system is established, which records the number of times the clamp head is used, the clamping time and the number of times it is disinfected, and displays a replacement prompt on the touch screen when the life threshold is reached.

[0014] Furthermore, the control unit integrates a data security and remote collaboration system to meet the needs of clinical medicine and teaching. Regarding data security, a three-tiered protection mechanism is adopted: during the data acquisition phase, sensor data is encrypted in real time; during the storage phase, hardware encryption is enabled on the SSD hard drive, and access permissions are set; during the transmission phase, SSL / TLS protocol encryption is used when transmitting data via Ethernet. The remote collaboration system supports two modes: one is real-time guidance, where the control unit connects to the cloud platform via a 5G module, allowing experts to view real-time surgical images and system parameters and send guidance instructions via remote terminals; the other is offline teaching, where the system automatically edits the surgical procedure into a teaching video, annotating key operational steps and parameter change curves.

[0015] Compared with existing technologies, the beneficial effects of this invention are: In terms of control precision and stability, the dynamic calibration step of the clamping force is corrected in real time by integrating multiple parameters such as temperature, air pressure, and time, effectively compensating for errors caused by environmental changes and mechanical fatigue, and ensuring that the clamping force is always maintained within the target range. Adaptive clamping employs a composite algorithm of fuzzy PID and feedforward compensation, which improves response speed and control stability, and can adapt to complex working conditions such as slight movement of the bleeding point, solving the problems of low control precision and weak anti-interference capability of traditional equipment. Multimodal fusion bleeding point localization technology, combined with endoscopy and near-infrared imaging, enhances the accuracy of bleeding point identification. Three-dimensional reconstruction and path planning functions ensure that the clamping tool accurately reaches the target position, avoiding operational deviations.

[0016] The safety protection mechanism is more comprehensive. The damage prevention protection steps calculate safety thresholds in layers and dynamically adjust the warning level based on parameters such as blood vessel diameter and clamping angle, achieving dual protection for both blood vessels and tissues and significantly reducing the risk of damage caused by excessive clamping. Hemostasis monitoring uses a multi-indicator fusion judgment, taking into account blood flow status, tissue oxygen supply, and tissue deformation, ensuring thorough hemostasis while avoiding tissue ischemia and hypoxia, thus balancing hemostasis effectiveness and tissue safety. Emergency intervention and fault self-repair functions can quickly respond to emergencies such as tool slippage and motor failure, shortening emergency response time and improving surgical safety through automatic handling and alarm prompts.

[0017] The system's adaptability and practicality have been significantly enhanced. The modular clamping actuator supports quick replacement of different types of clamp heads, adapting to the needs of various surgical sites such as coronary arteries and peripheral vessels. Redundant design and intelligent maintenance mechanisms extend the equipment's lifespan and reduce maintenance costs. The control unit integrates data security and remote collaboration functions, complying with medical data privacy regulations while enabling remote expert guidance and surgical teaching, expanding the equipment's application scenarios. The human-machine interface unit is easy to operate, supporting parameter visualization and manual intervention, conforming to clinical operating habits and improving the doctor's user experience.

[0018] Overall, this invention achieves precise, intelligent, and safe control of hemostasis clamping in minimally invasive cardiovascular surgery, effectively reducing the risk of surgical complications, alleviating the workload of doctors, and improving the clinical adaptability and reliability of the equipment. It provides strong support for the safe implementation of minimally invasive cardiovascular surgery and has broad prospects for clinical promotion. Attached Figure Description

[0019] Figure 1 This is a schematic block diagram of the hemostasis clamping control system for cardiovascular minimally invasive surgery proposed in this invention. Figure 2 This is a schematic block diagram of the hemostasis clamping control method for cardiovascular minimally invasive surgery proposed in this invention; Figure 3 A comparison chart of the clamping force control accuracy of different hemostasis methods; Figure 4 A comparison chart of bleeding point identification accuracy versus identification time; Figure 5 A comparison chart of security performance indicators of different systems. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0022] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances. The invention will now be described in further detail with reference to the accompanying drawings.

[0023] Reference Figures 1 to 5 A hemostatic clamping control method for minimally invasive cardiovascular surgery includes the following steps: Step 1: Bleeding Point Localization and Feature Recognition. Surgical field images were acquired using a 4K endoscope with a resolution of 3840×2160 and a frame rate of 60fps. The U-Net++ segmentation model was used to process the acquired images to extract the bleeding point region; this model has an mIOU of at least 0.95. The bleeding type was identified by combining HOG features with an SVM classifier, which identified arterial spurting, venous oozing, and capillary oozing. Simultaneously, a laser rangefinder was used to measure the distance from the bleeding point to the clamping tool; this sensor had an accuracy of ±0.1mm.

[0024] Step 2: Initialize clamping parameters. Preset the clamping force baseline value based on the identified bleeding type: 3-5N for arterial bleeding, 1-2N for venous bleeding, and 0.5-1N for capillary bleeding. Set the initial opening degree of the clamping tool using a displacement sensor with a range of 8-12mm (0-50mm range, ±0.01mm accuracy). Adjust the clamping angle based on the blood vessel diameter at the surgical site measured via endoscopic images; the clamping angle is adjustable within the range of 0-30°.

[0025] Step 3: Adaptive Clamping Control. Start the electric clamping tool, which is driven by a stepper motor with a torque accuracy of ±0.01. The clamping tool is controlled to approach the bleeding point according to preset parameters. The actual clamping force is fed back in real time by a pressure sensor installed on the clamping tool. The pressure sensor has a range of 0-10N and an accuracy of ±0.02N. The actual clamping force is compared with a preset reference value, and the motor speed is dynamically adjusted according to the comparison result. The motor speed range is 50-300rpm to ensure that the clamping force error does not exceed 0.1N.

[0026] Step 4: Monitoring Hemostasis Effectiveness. The blood flow status at the bleeding point is analyzed using endoscopic images, specifically by calculating blood flow velocity using optical flow. A pulse oximeter sensor with a sampling rate of 10Hz and an accuracy of ±1% is used to monitor the oxygen supply to the local tissue. The criteria for successful hemostasis are cessation of bleeding and tissue oxygenation not falling below 90%.

[0027] Step 5: Clamping and Release. After successful hemostasis, the control device switches to constant force maintenance mode, using a PID controller to compensate for the mechanical drift of the clamping tool, ensuring the drift error does not exceed 0.05N. After the surgery, the clamping force is reduced at a preset rate gradient of 0.5-1N / s until the clamping tool is completely released to avoid secondary damage to the blood vessel.

[0028] This invention also includes a dynamic clamping force calibration step, which is integrated throughout the entire process of adaptive clamping control and constant force maintenance. This step improves control accuracy by collaboratively correcting clamping force deviations through multiple parameters. First, surgical environment parameters are collected in real time using temperature and pressure sensors. The temperature measurement range is 18-30℃ with an accuracy of ±0.2℃, and the pressure measurement range is 95-105kPa with an accuracy of ±0.1kPa. Simultaneously, the continuous working time of the clamping tool is recorded, with the time recording accurate to 0.1s. Based on the collected parameters, the calibrated clamping force is calculated using the following formula:

[0029] in, The clamping force after calibration is expressed in N. This is a preset clamping force reference value, in N; This is a temperature coefficient, with a value of -0.005 / ℃, used to correct for changes in the material's elastic modulus caused by temperature variations. This represents the difference between the actual surgical environment temperature and the standard temperature, where the standard temperature is 25℃, and the unit is ℃. This is the air pressure coefficient, with a value of 0.002 / kPa, used to compensate for the influence of air pressure on the sensor's detection accuracy; This is the difference between the actual air pressure and the standard air pressure, which is 101.3 kPa. The unit is kPa. The time decay coefficient, with a value of -0.001 / s, is used to correct for mechanical fatigue errors after prolonged use of the clamping tool; t is the clamping duration in seconds. The calibration frequency is 10Hz. After each calibration, the motor output torque is adjusted via a PWM signal to ensure that the actual clamping force remains stable. Within the range.

[0030] This invention also includes a damage prevention protection step, which is initiated after the clamping parameters are initialized. This step achieves dual protection for both blood vessels and tissues through layered calculation of safety thresholds. First, the diameter of the blood vessel at the surgical site is measured using endoscopic images, with a measurement range of 1-10 mm and an accuracy of ±0.1 mm. The clamping contact area is calculated by combining the jaw dimensions of the clamping tool (width 5-10 mm, thickness 0.5-1 mm) with the set opening degree. Then, the maximum compressive strength is determined according to the blood vessel type; for arteries, 3 × 10⁻⁶ is used. Pa, 1× intravenous injection Pa, capillary blood sample 5× Pa. The safety clamping force threshold is calculated using the following formula:

[0031] in, The safety clamping force threshold is expressed in nanometers (N). The maximum compressive strength of the blood vessel at the surgical site is expressed in Pa; A is the contact area between the clamping tool and the blood vessel, expressed in m². The clamping angle is expressed in rad. The correction factor for vessel diameter is 0.02 / mm, with a larger correction range for smaller diameter vessels; D is the vessel diameter in mm. A three-level early warning mechanism is set: when the actual clamping force reaches... When the pressure reaches 70%, the console displays a yellow warning; when it reaches 80%, the clamping force increase rate is reduced to 0.05 N / s; when it reaches 90%, a red warning is issued and the clamping force is locked, allowing adjustment only after manual confirmation by the doctor, to avoid excessive compression of the blood vessel leading to intimal damage or rupture.

[0032] In this invention, step 1, bleeding point localization employs multimodal fusion and 3D reconstruction technology to improve localization accuracy. First, 4K endoscopic images and near-infrared imaging data are simultaneously acquired. The near-infrared light source wavelength is 700-900nm. A narrow-band filter filters stray light, enhancing the grayscale difference between hemoglobin and surrounding tissue, resulting in a bleeding point contrast improvement of at least 40%. The two images are input into an image fusion module, employing a Laplacian pyramid fusion algorithm. The bottom layer retains detailed information from the endoscopic image, while the top layer overlays the bleeding point features from the near-infrared image. The fused image resolution remains 3840×2160. The fused stereo image pair is acquired through a binocular vision system. The baseline distance between the left and right cameras of the binocular vision system is 10mm, with a synchronization frame rate of 60fps. The disparity map is calculated, and the 3D coordinates of the bleeding point are reconstructed, with a coordinate error not exceeding 0.2mm. Combined with the surgical robot's robotic arm coordinate system, which uses an absolute encoder with a positioning accuracy of ±0.1mm, the bleeding point coordinates are converted into target points in the robotic arm's base coordinate system, generating the motion path of the gripping tool. Path planning utilizes… The algorithm avoids important surrounding tissues such as the heart and nerve bundles, with a path smoothness error of ≤0.5mm.

[0033] In this invention, the adaptive clamping control in step 3 employs a composite control algorithm combining fuzzy PID and feedforward compensation. First, the clamping force deviation is defined. ( (actual clamping force) and rate of change of deviation The fuzzy sets of both are divided into seven levels: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large. The membership function adopts a triangular distribution. A fuzzy rule base containing 25 rules is established, such as "if e is positive large and ec is negative small, then the proportional coefficient is large, the integral coefficient is small, and the derivative coefficient is medium." Fuzzy reasoning and the centroid method are used to defuzzify the PID parameters: the proportional coefficient ranges from 0.1 to 1.0, the integral coefficient ranges from 0.01 to 0.1, and the derivative coefficient ranges from 0.001 to 0.01. At the same time, a feedforward compensation term is introduced to calculate the required pre-torque of the motor in advance based on the distance change rate measured by the laser rangefinder, compensating for the lag caused by motion inertia. This composite algorithm ensures that the control response time does not exceed 0.5s, the overshoot does not exceed 5%, and the clamping force remains stable with an error ≤0.08N even when the bleeding point position shifts slightly by ≤1mm.

[0034] In this invention, step 4, hemostasis effect monitoring, employs multi-indicator fusion and dynamic judgment logic to ensure a balance between hemostasis effect and tissue safety. Four core monitoring indicators are set: first, blood flow velocity, calculated using optical flow to measure pixel movement velocity in the bleeding area, with a threshold of ≤0.5cm / s; second, bleeding area, segmented in real-time using a U-Net++ model to calculate the area reduction ratio, with a threshold of ≥95%; third, local oxygen saturation, monitored by a fiber optic oxygen probe (0.8mm diameter, 2-3mm insertion depth), with a threshold of ≥92%; and fourth, tissue deformation at the clamping site, analyzed using endoscopic images to assess the degree of vessel wall indentation, with a threshold of ≤0.2mm. A dynamic judgment rule is established: In the initial stage of surgery (0-3 seconds after clamping), initial hemostasis is determined when blood flow velocity is ≤0.5cm / s and bleeding area is reduced by ≥90%; in the stable stage (3-10 seconds), blood oxygen saturation must be ≥92% and tissue deformation ≤0.2mm, otherwise it is judged as excessive clamping; in the sustained stage (more than 10 seconds), at least 2 of the four indicators must be met simultaneously and the result must be sustained for more than 3 seconds to determine successful hemostasis. If the criteria are not met, the system automatically initiates an adjustment procedure: first, increase the clamping force by 0.2N, maintain for 1 second, and then check the indicators; if the criteria are still not met, adjust the clamping angle by 5° and increase the clamping force by 0.2N again, up to 3 times. If the adjustment still fails, a manual intervention prompt is triggered.

[0035] This invention also includes emergency intervention and self-repair steps to improve the system's reliability in emergency situations. The system monitors key parameters in real time: clamping force sampling frequency 100Hz, bleeding velocity sampling frequency 20Hz, and motor current sampling frequency 50Hz. Three types of emergency events are defined: first, tool slippage, determined by a sudden drop in clamping force of ≥1N within 0.1s and no significant change in bleeding velocity; second, secondary bleeding, determined by a sudden increase in bleeding velocity of ≥2cm / s within 0.5s and stable clamping force; and third, motor failure, determined by a motor current exceeding the rated value of 1.5A and no response in clamping force. In the event of an emergency, a three-tiered response is immediately initiated: Level 1 braking, where the motor stops moving within 0.1 seconds and maintains the current clamping state; Level 2 alarm, where a buzzer sounds continuously at a frequency of 2kHz, a red indicator light on the surgical console flashes, and the touchscreen displays the fault type and suggested actions; Level 3 self-repair, where for tool slippage, the system automatically increases the clamping force slightly by 0.3-0.5N and fine-tunes the angle by 3° to attempt re-clamping; for slight motor jamming, a reverse pulse signal is output for 0.05 seconds to release the jamming, and if repair fails, the motor is locked and awaits manual intervention. All operational data is stored in real time during emergency intervention for easy postoperative review and analysis.

[0036] This invention includes the following modules: The image acquisition and recognition unit consists of a 4K endoscope, a near-infrared imaging module, a binocular vision assembly, and an image processor. The 4K endoscope uses a CMOS image sensor with a pixel size of 1.55μm, equipped with a 12x optical zoom lens with a focal length of 5-60mm and an aperture of F1.8-F2.8, achieving clear imaging within a 0.5-10cm range via motorized focusing. The near-infrared imaging module includes a 700-900nm wavelength LED light source with a power of 100mW and a narrowband filter. The left and right cameras of the binocular vision assembly are triggered synchronously to ensure consistent image acquisition timing. The image processor uses the NVIDIA Jetson AGX Orin platform with a GPU computing power of at least 200 TOPS, capable of running the U-Net++ segmentation model and image fusion algorithm in parallel with a processing latency of ≤50ms.

[0037] The clamping unit consists of a modular electric clamping jaw, a stepper motor drive module, and a displacement detection component. The clamping jaws are forged from medical-grade TC4 titanium alloy and coated with PTFE (polytetrafluoroethylene) with a thickness of 20μm and a friction coefficient ≤0.15. They are available in straight and bent head versions; the straight head has a jaw width of 5mm, while the bent head has a 30° bend and an 8mm jaw width. They connect to the clamp body via a quick-change interface, with a replacement time of no more than 10 seconds. The stepper motor is a two-phase hybrid type with a step angle of 1.8°, achieving 0.09° micro-step control through a microstepping driver, and an output torque range of 0.1-0.5. The displacement detection component uses a magnetic scale with a range of 0-50mm and an accuracy of ±0.01mm, providing real-time feedback on the jaw opening degree.

[0038] The sensing feedback unit integrates a pressure sensor, a blood oxygen saturation sensor, a temperature and pressure sensor, and a current sensor. The pressure sensor is a thin-film piezoresistive sensor, mounted inside the jaws, measuring 5×3mm, with a range of 0-10N, an accuracy of ±0.02N, and outputting a 4-20mA analog signal. The blood oxygen saturation sensor is a fiber optic probe type with a probe diameter of 0.8mm, measured using spectroscopy, employing 660nm red light and 940nm infrared light, a sampling rate of 10Hz, and an accuracy of ±1%. The temperature and pressure sensors use an integrated chip, simultaneously outputting temperature and pressure data; the temperature measurement range is 18-30℃ with an accuracy of ±0.2℃, and the pressure measurement range is 95-105kPa with an accuracy of ±0.1kPa. The current sensor is connected in series in the motor power supply circuit, with a range of 0-5A and an accuracy of ±0.01A, monitoring the motor's operating status.

[0039] The control unit adopts an FPGA+ARM heterogeneous architecture. The FPGA is a Xilinx XC7K325T, responsible for image preprocessing, sensor signal acquisition, and motor pulse output, with a processing speed ≤1μs. The ARM is an STM32H743 with a main frequency of 480MHz, running the fuzzy PID algorithm, emergency intervention logic, and data storage program. The control unit has built-in 1GB DDR4 memory and a 1TB SSD hard drive, supporting real-time data caching and long-term storage. It communicates with other units via a CAN bus with a bus rate of 1Mbps, and the control command output latency is ≤10ms.

[0040] The human-machine interface unit includes a 10.1-inch multi-touch LCD screen, a foot switch, and an audible and visual alarm device. The touchscreen has a resolution of 1920×1080 and a brightness of 500 cd / m², and can display surgical field images, clamping force curves, sensor parameters, and system status. It supports doctors in manually setting parameters such as clamping force and opening / closing degree. The foot switch has three positions: clamping, releasing, and fine adjustment. It is waterproof and has a response time of ≤0.1s. The audible and visual alarm device includes a buzzer and red, yellow, and green indicator lights. The buzzer frequency is 2kHz, and the three colors of the lights correspond to emergency faults, warnings, and normal status, respectively.

[0041] In this invention, the clamping execution unit adopts a redundant design and intelligent maintenance mechanism to improve long-term operational reliability. The motor drive module has a built-in dual-path drive circuit. When the main circuit fails, the backup circuit automatically switches within 0.2 seconds to ensure uninterrupted clamping action. An automatic grease replenishment device is installed inside the clamp body, automatically squeezing out 0.1mL of medical-grade grease every 10 hours of motor operation to maintain the flexibility of the transmission components. A clamp head life management system is established, recording the number of times the clamp head is used, clamping time, and disinfection times. The maximum number of uses is 50 times. When the life threshold is reached, the touch screen displays a replacement prompt. The motor housing is made of die-cast aluminum alloy with an anodized surface treatment, possessing an IP67 waterproof rating and corrosion resistance, suitable for disinfectant spray environments during surgery. In addition, the clamping execution unit is equipped with calibration fixtures. Before each use, the pressure sensor is calibrated with standard weights ranging from 0.1-10N to ensure a detection accuracy error of ≤0.02N.

[0042] In this invention, the control unit integrates a data security and remote collaboration system to meet the needs of clinical medicine and teaching. Regarding data security, a three-tiered protection mechanism is adopted: during the data acquisition phase, sensor data is encrypted in real time using the AES-256 encryption algorithm; during the storage phase, the SSD hard drive is hardware encrypted with access permissions set: doctors can read and write, nurses can read only, and visitors can view only; during the transmission phase, data is encrypted using the SSL / TLS protocol at a 1Gbps Ethernet speed to prevent data leakage. The remote collaboration system supports two modes: one is real-time guidance, where the control unit connects to the cloud platform via a 5G module that supports SA (Standalone) networking, allowing experts to view real-time surgical images and system parameters and send guidance commands via remote terminals with a latency of ≤200ms; the other is offline teaching, where the system automatically edits the surgical procedure into a teaching video, annotating key operation steps and parameter change curves. Key steps include clamping force adjustment and angle optimization, and support for speed playback and resume playback. Furthermore, the control unit has a reserved DICOM interface for integration with the hospital's PACS system, enabling the associated storage of surgical data and patient medical records, complying with medical information technology standards.

[0043] The following two examples further illustrate the specific implementation of this system: Example 1: Hemostasis and control of arterial bleeding during coronary interventional surgery This embodiment addresses the hemostasis and control of arterial spurting bleeding (approximately 2.5 mm in diameter) that occurs during stent implantation for left anterior descending coronary artery stenosis. The specific procedure is as follows: During the preoperative preparation phase, the straight-tipped clamp is installed onto the end of the surgical robotic arm via a quick-change interface. The clamp head is forged from medical-grade TC4 titanium alloy with a 20μm thick polytetrafluoroethylene coating and a 5mm jaw width. The system self-test is initiated, and the control unit calibrates the pressure sensor using standard weights (0.5N, 1N, 3N, 5N) to ensure a detection error ≤0.02N. The laser rangefinder sensor achieves an accuracy of ±0.1mm after calibration. The 4K endoscope and near-infrared imaging module are simultaneously activated, adjusted to a resolution of 3840×2160 and a frame rate of 60fps, with the near-infrared light wavelength set to 780nm and stray light filtered through a narrow-band filter.

[0044] Step 1: Bleeding Point Localization and Feature Recognition. The endoscope acquires images of the puncture area. The U-Net++ segmentation model extracts the bleeding point region within 0.3 seconds, achieving an mIOU of 0.96 and identifying jet-like blood flow characteristics. The HOG feature extractor captures the blood flow direction vector, and the SVM classifier determines it as arterial spurting bleeding based on the 128-dimensional feature vector. The laser rangefinder takes 10 consecutive samples and averages the results, measuring the distance from the bleeding point to the clamping forceps at 8.3 mm. Near-infrared imaging shows that the grayscale value of the hemoglobin region at the bleeding point is 42% higher than the surrounding tissue. The Laplacian pyramid fusion algorithm overlays the endoscopic image details with the near-infrared features to generate a fused image. The binocular vision system (baseline distance 10 mm) acquires stereo image pairs. The 3D coordinates of the bleeding point (X=14.2 mm, Y=9.7 mm, Z=6.1 mm) are reconstructed through parallax calculation. The coordinate transformation module maps these coordinates to the robotic arm's base coordinate system. The algorithm plans a motion path that avoids the left ventricle and circumflex artery vessels, with a total path length of 23 mm and a smoothness error of 0.3 mm.

[0045] Step 2: Clamping Parameter Initialization. Based on the arterial bleeding type, the system retrieves baseline parameters from the database: a baseline clamping force of 4N and an initial opening / closing angle of 10mm. The endoscopic image measurement module measures the vessel diameter as 2.5mm using an edge detection algorithm. The control unit calculates the clamping angle correction value according to the formula and finally sets the angle to 15°. The displacement sensor monitors the clamping jaw position in real time to ensure the opening / closing angle remains stable at 10mm ± 0.01mm.

[0046] Step 3: Adaptive Clamping Control. The control unit sends motion commands to the stepper motor drive module. The motor drives the clamping jaws to move along the planned path at a speed of 200 rpm, reaching the target position after 1.2 seconds. The pressure sensor begins to provide feedback on the actual clamping force, with an initial value of 0.3 N. The system calculates the clamping force deviation e = 3.7 N and the deviation change rate ec = 2.1 N / s. The fuzzy PID controller calls rule number 18 from the rule base (e = positive, ec = positive), outputting a proportional coefficient of 0.6, an integral coefficient of 0.05, and a derivative coefficient of 0.005. Simultaneously, the feedforward compensation module calculates a pre-torque of 0.03 based on the distance change rate (0.8 mm / s) measured by the laser rangefinder. The clamping force is applied to the motor in advance. When the clamping force reaches 3.8N, the system initiates dynamic calibration. The actual temperature was 26℃, the actual air pressure was 102kPa, and t was the clamping duration. Calibration was performed every 0.1s, adjusting the motor output via a PWM signal. After 3s, the clamping force stabilized within the range of 4N±0.04N.

[0047] Step 4: Monitoring Hemostasis Effectiveness. Optical flow analysis was used to analyze the motion trajectory of 200 feature points in the bleeding area, calculating the blood flow velocity as it decreased from an initial 1.5 cm / s to 0.3 cm / s. The U-Net++ model segmented the bleeding area in real time, reducing the area from 8 mm² to 0.2 mm², a reduction of 97.5%. A fiber optic pulse oximeter (0.8 mm diameter, 2.5 mm insertion depth) monitored local oxygen saturation at 94%. The image deformation analysis module tracked the vessel wall edge, measuring a maximum indentation of 0.15 mm. Successful hemostasis was determined when all four indicators met the following criteria: blood flow velocity ≤ 0.5 cm / s, bleeding area reduction ≥ 95%, and oxygen saturation ≥ 92%, sustained for 3.2 seconds.

[0048] Step 5: Clamping and Release. The control unit switches to constant force maintenance mode, and the PID controller parameters are adjusted to proportional 0.4, integral 0.1, and derivative 0.002, keeping the mechanical drift within 0.03N. The surgery is completed after 15 minutes. The system gradually reduces the clamping force at a rate of 0.8N / s, with the force changing from 4N→3.2N→2.4N→1.6N→0.8N→0, over a total of 30 seconds. Bleeding is monitored in real-time during release, and no secondary bleeding occurs.

[0049] The damage protection system operates synchronously and calculates the safety threshold: The actual clamping force of 4N was 75.5% of the threshold, which only triggered a level one yellow warning and did not affect normal operation.

[0050] Example 2: Hemostasis and control of venous oozing after femoral artery puncture This embodiment is used to treat venous bleeding with a diameter of 5mm after femoral artery puncture. The procedure is as follows: Preoperatively, the clamps were replaced with 30° bent-head clamps (jaw width 8mm), with a replacement time of 8 seconds. The system automatically identified the clamp head type and loaded the corresponding parameter library. The temperature and pressure sensors displayed the following environmental parameters: temperature 24℃, pressure 100.5kPa.

[0051] Step 1: Bleeding point localization and feature recognition. Images were simultaneously acquired using a 4K endoscope and 850nm near-infrared imaging. The bleeding area showed a significant grayscale difference in the near-infrared image, and the contrast was improved by 42% after fusion. The U-Net++ model extracted the bleeding area as 15mm², and the SVM classifier determined it to be venous bleeding. The binocular vision system measured the three-dimensional coordinates of the bleeding point and planned the path, avoiding branches of the femoral nerve and femoral vein.

[0052] Step 2: Initialize clamping parameters. Based on the type of venous bleeding, preset the clamping force baseline value to 1.5N and the initial opening / closing degree to 12mm. The measured blood vessel diameter is 5mm, and the clamping angle is adjusted to 20°.

[0053] Step 3: Adaptive clamping control. A stepper motor drives the clamping jaws to move at 150 rpm, and the pressure sensor feedback force reaches 1.5 N ± 0.03 N. Dynamic calibration formula: After compensation, the stability of the force value is improved by 30%.

[0054] Step 4: Monitoring hemostasis effect. Multiple indicators showed a blood flow velocity of 0.4 cm / s, a 96% reduction in bleeding area, a blood oxygen saturation of 93%, and a tissue deformation of 0.1 mm, meeting the hemostasis criteria.

[0055] Step 5: Holding and releasing. During the constant force maintenance phase, the drift error was 0.02N, the release rate was 0.6N / s, and the device was fully released after 40s without any vasospasm.

[0056] Performance data representation

[0057] Table 1 Table 1 shows the data from statistical results of 50 clinical simulation experiments (25 coronary artery surgeries and 25 peripheral vascular surgeries). The significant reduction in clamping force control error is attributed to the comprehensive compensation of temperature, air pressure, and time factors by the dynamic calibration formula, as well as the synergistic control of fuzzy PID and feedforward compensation, achieving sub-Newtonian level force accuracy. The hemostasis success rate increased to 98%, mainly due to the multimodal fusion bleeding point recognition technology (97% accuracy) and multi-index judgment logic, ensuring precise matching of clamping parameters to bleeding types. The incidence of vascular injury decreased to 2%, validating the effectiveness of the damage prevention system. Dynamic adjustment of the clamping upper limit through a safety threshold formula avoided endovascular damage caused by excessive compression. The significant reduction in surgical time reflects the fully automated design of the system from positioning to release, particularly the clamp head replacement completed within 10 seconds and the emergency response within <0.5 seconds, significantly improving surgical efficiency. The multimodal recognition accuracy is 12 percentage points higher than existing equipment, demonstrating the advantages of near-infrared fusion and deep learning algorithms, effectively solving the recognition challenges of blood reflection and tissue occlusion. Overall data shows that the technology of this invention has achieved a qualitative leap in control precision, safety and operational efficiency, meeting the stringent requirements of cardiovascular minimally invasive surgery.

[0058] refer to Figure 3This figure demonstrates the significant advantages of the present invention in terms of clamping force control precision. Traditional manual operation is greatly affected by the doctor's feel and the environment, with an average error of ±0.54N, which can easily lead to incomplete hemostasis or vascular damage. Although existing automated equipment has improved this, the average error is still ±0.35N, and the precision fluctuates significantly with changes in the environment. The present invention compensates for errors caused by temperature, air pressure, and time through a dynamic calibration formula, combined with a fuzzy PID algorithm, reducing the average error to ±0.05N, and exhibiting small error fluctuations under different environmental conditions. This ensures that the clamping force remains stable within the target range, providing a core guarantee for precise hemostasis.

[0059] refer to Figure 4 This figure highlights the superiority of the multimodal fusion recognition technology of this invention. Single-endoscope recognition is affected by blood reflection and tissue occlusion, resulting in low accuracy and slow recognition speed. While existing multimodal recognition methods have improved, their performance in capillary bleeding and occluded scenarios remains unsatisfactory. This invention fuses 4K endoscopy with 700-900nm near-infrared imaging, combining the U-Net++ segmentation model and SVM classifier. The accuracy rates for arterial, venous, and capillary bleeding reach 98%, 96%, and 92%, respectively, with an accuracy rate of 89% in occluded scenarios. Furthermore, the average recognition time is reduced to 280ms. This rapid and accurate recognition allows more time for subsequent clamping operations, improving surgical efficiency.

[0060] refer to Figure 5 This radar chart reflects the comprehensive safety advantages of the system of this invention. Traditional manual systems lack early warning and rapid response capabilities, resulting in a high incidence of vascular injury and tissue ischemia; while existing automated systems possess basic safety functions, their early warning accuracy and response speed are insufficient. This invention calculates safety thresholds using an anti-damage protection formula, achieving an over-clamping early warning accuracy of 98% and reducing the incidence of vascular injury to 2%; multi-indicator monitoring enables a secondary bleeding detection rate of 99% and a tissue ischemia incidence of 1%; the emergency intervention module achieves a response time of <500ms, significantly reducing surgical risks and providing multiple safeguards for patient safety.

[0061] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A hemostatic clamping control method for minimally invasive cardiovascular surgery, characterized in that, Includes the following steps: Step 1: Bleeding Point Localization and Feature Recognition Steps: Surgical field images are acquired using a 4K endoscope. The acquired images are processed using the U-Net++ segmentation model to extract the bleeding point region. The bleeding type is identified by combining HOG features and an SVM classifier. At the same time, the distance from the bleeding point to the clamping tool is measured using a laser rangefinder. Step 2: Clamping parameter initialization steps: Preset the clamping force reference value according to the identified bleeding type; set the initial opening degree of the clamping tool through the displacement sensor; Adjust the clamping angle based on the diameter of the blood vessel at the surgical site obtained through endoscopic images; Step 3: Adaptive clamping control steps: Start the electric clamping tool, which is driven by a stepper motor; control the clamping tool to approach the bleeding point according to the preset parameters, and provide real-time feedback of the actual clamping force through the pressure sensor installed on the clamping tool; compare the actual clamping force with the preset reference value, and dynamically adjust the motor speed according to the comparison result; Step 4: Monitoring the hemostasis effect: Analyze the blood flow status of the bleeding point through endoscopic images, specifically using optical flow to calculate blood flow velocity; and monitor the oxygen supply to the local tissue using a blood oxygen saturation sensor. Step 5: Clamping State Maintenance and Release Steps: After successful hemostasis, the control device switches to constant force maintenance mode and compensates for the mechanical drift of the clamping tool through the PID controller; after the operation is completed, the clamping force is reduced according to the preset rate gradient until the clamping tool is completely released.

2. The hemostatic clamping control method for cardiovascular minimally invasive surgery according to claim 1, characterized in that, Also includes: The clamping force dynamic calibration step is carried out throughout the entire process of adaptive clamping control and constant force maintenance. It improves control accuracy by coordinating multiple parameters to correct clamping force deviation. First, surgical environment parameters are collected in real time using temperature and pressure sensors, and the continuous working time of the clamping tool is recorded. After calibration, the clamping force calculation formula is as follows: ,in, To calibrate the clamping force, To preset the clamping force reference value, Temperature coefficient; This represents the difference between the actual surgical environment temperature and the standard temperature. This is the difference between the actual air pressure and the standard air pressure; after each calibration, the motor output torque is adjusted via a PWM signal.

3. The hemostatic clamping control method for minimally invasive cardiovascular surgery according to claim 1, characterized in that, Also includes: The damage prevention and protection step is initiated after the clamping parameters are initialized. It achieves dual protection of blood vessels and tissues by calculating the safety threshold in layers. First, the diameter of the blood vessel at the surgical site is measured by endoscopic images, the clamping contact area is calculated, and then the maximum compressive strength is determined according to the type of blood vessel. The safety clamping force threshold is calculated using the following formula: ,in, For the safety clamping force threshold, A represents the maximum compressive strength of the blood vessel at the surgical site, and A is the contact area between the clamping tool and the blood vessel. For the clamping angle, D is the blood vessel diameter correction factor; D is the blood vessel diameter; a three-level early warning mechanism is set: when the actual clamping force reaches... When the clamping force reaches 70%, the control panel displays a yellow warning; when it reaches 80%, the clamping force growth rate is reduced to 0.05 N / s; when it reaches 90%, a red warning is issued and the clamping force is locked.

4. The hemostatic clamping control method for minimally invasive cardiovascular surgery according to claim 1, characterized in that, In step 1, the bleeding point localization adopts multimodal fusion and three-dimensional reconstruction technology to improve the localization accuracy. First, 4K endoscopic images and near-infrared imaging data are acquired simultaneously. Narrow-band filters are used to filter stray light and enhance the grayscale difference between hemoglobin and surrounding tissues. The two images are input into the image fusion module, and the Laplacian pyramid fusion algorithm is used. The bottom layer retains the detailed information of the endoscopic image, and the top layer is superimposed with the bleeding point features of the near-infrared image. The system acquires fused stereo image pairs through a binocular vision system, calculates the disparity map, and reconstructs the three-dimensional coordinates of the bleeding point. Combined with the robotic arm coordinate system of the surgical robot, which uses an absolute encoder, the coordinates of the bleeding point are converted into target points in the robotic arm base coordinate system, generating the motion path of the gripping tool.

5. The hemostatic clamping control method for minimally invasive cardiovascular surgery according to claim 1, characterized in that, In step 3, the adaptive clamping control employs a composite control algorithm combining fuzzy PID and feedforward compensation; firstly, the clamping force deviation is defined. and the rate of change of deviation ,in To determine the actual clamping force, the fuzzy sets of both are divided into seven levels: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large. The membership function adopts a triangular distribution. A fuzzy rule base containing 25 rules is established, and a feedforward compensation term is introduced. Based on the distance change rate measured by the laser rangefinder, the pre-torque required by the motor is calculated in advance to compensate for the lag caused by motion inertia.

6. The hemostatic clamping control method for cardiovascular minimally invasive surgery according to claim 1, characterized in that, Step 4 monitors the hemostasis effect using a multi-indicator fusion and dynamic judgment logic; four core monitoring indicators are set: first, blood flow velocity, which is calculated by optical flow method to determine the pixel movement velocity of the bleeding point area; second, bleeding area, which is calculated by real-time segmentation of the bleeding area using the U-Net++ model to determine the area reduction ratio. Thirdly, local blood oxygen saturation is monitored by using a fiber optic blood oxygen probe to monitor the tissues surrounding the clamping site. Fourth, tissue deformation at the clamping site is analyzed using endoscopic images to determine the degree of indentation in the vessel wall. A dynamic judgment rule is established: In the initial stage of surgery, initial hemostasis is determined when blood flow velocity is ≤0.5cm / s and bleeding area is reduced by ≥90%. In the stable stage, blood oxygen saturation must be ≥92% and tissue deformation ≤0.2mm; otherwise, it is considered excessive clamping. In the sustained stage, at least two of the four indicators must be met simultaneously and maintained for more than 3 seconds to determine successful hemostasis. If the criteria are not met, the system automatically initiates an adjustment procedure: first, the clamping force is increased by 0.2N, maintained for 1 second, and then the indicators are checked. If the criteria are still not met, the clamping angle is adjusted by 5° and the clamping force is increased by 0.2N again, with a maximum of 3 adjustments. If the procedure still fails, a manual intervention prompt is triggered.

7. The hemostatic clamping control method for cardiovascular minimally invasive surgery according to claim 1, characterized in that, It also includes emergency intervention and fault self-repair steps, defining three types of emergency events: tool slippage, secondary bleeding, and motor failure. When an emergency occurs, a three-level response is immediately initiated: Level 1 braking, the motor stops moving within 0.1 seconds and maintains the current clamping state; Level 2 alarm, the buzzer sounds continuously at a frequency of 2kHz, the red indicator light on the surgical console flashes, and the fault type and handling suggestions are displayed on the touch screen. The system has three levels of self-repair. In case of tool slippage, the system automatically increases the clamping force slightly by 0.3-0.5N and fine-tunes the angle by 3° to attempt to re-clamp. In case of slight motor jamming, the system outputs a reverse pulse signal for 0.05 seconds to release the jamming. If the repair fails, the system locks the motor and waits for manual intervention.

8. A system for implementing the hemostatic clamping control method for cardiovascular minimally invasive surgery according to any one of claims 1-7, comprising: The image acquisition and recognition unit consists of a 4K endoscope, a near-infrared imaging module, a binocular vision component, and an image processor. The 4K endoscope uses a CMOS image sensor. The near-infrared imaging module contains a 700-900nm LED light source. The left and right cameras of the binocular vision component are triggered synchronously. The image processor uses the NVIDIA Jetson AGX Orin platform. The clamping execution unit consists of a modular electric clamping clamp, a stepper motor drive module, and a displacement detection component. The sensing feedback unit integrates a pressure sensor, a blood oxygen saturation sensor, a temperature and pressure sensor, and a current sensor; the blood oxygen saturation sensor is a fiber optic probe type, measured by spectroscopy; the temperature and pressure sensor uses an integrated chip to simultaneously output temperature and pressure data to monitor the motor's operating status. The control unit is responsible for image preprocessing, sensor signal acquisition, and motor pulse output; the control unit supports real-time data caching and long-term storage. The human-computer interaction unit displays surgical field images, clamping force curves, sensor parameters, and system status, and supports doctors to manually set parameters such as clamping force and opening degree; the foot switch has three positions: clamping, releasing, and fine adjustment; the audible and visual alarm device includes a buzzer and red, yellow, and green indicator lights, with the buzzer frequency at 2kHz and the three-color lights corresponding to emergency fault, warning, and normal status, respectively.

9. The system for hemostasis clamping control method in cardiovascular minimally invasive surgery according to claim 8, characterized in that, The clamping execution unit adopts a redundant design and intelligent maintenance mechanism; the motor drive module has a built-in dual-drive circuit, and the backup circuit automatically switches within 0.2 seconds when the main circuit fails; the clamp body is equipped with an automatic grease replenishment device to maintain the flexibility of the transmission components; a clamp head life management system is established, which records the number of times the clamp head is used, the clamping time and the number of times it is disinfected, and displays a replacement prompt on the touch screen when the life threshold is reached.

10. The system for hemostasis clamping control method in cardiovascular minimally invasive surgery according to claim 8, characterized in that, The control unit integrates a data security and remote collaboration system to meet the needs of clinical medicine and teaching. For data security, a three-tiered protection mechanism is employed: during data acquisition, sensor data is encrypted in real time; during storage, hardware encryption is enabled on the SSD, and access permissions are set; during transmission, SSL / TLS encryption is used when transmitting data over Ethernet. The remote collaboration system supports two modes: real-time guidance, where the control unit connects to the cloud platform via a 5G module, allowing experts to view real-time surgical images and system parameters and send guidance instructions via remote terminals; and offline teaching, where the system automatically edits the surgical procedure into a teaching video, annotating key operational steps and parameter change curves.

Images