CN-121987351-A - Intelligent control method and system for microsurgery robot

CN121987351ACN 121987351 ACN121987351 ACN 121987351ACN-121987351-A

Abstract

The application provides an intelligent control method and system for a microsurgery robot, and belongs to the technical field of medical robot control. The method comprises the steps of obtaining visual information and multidimensional force information aiming at a surgical instrument when each surgical step is executed based on a step time sequence library, calculating visual deviation of an instrument tip based on the visual information and calculating force abnormality indexes based on the multidimensional force information, judging whether preset abnormality triggering conditions are met or not based on the visual deviation and the force abnormality indexes by taking the currently executed surgical step as a decision context, if yes, suspending the currently executed surgical step, determining a deviation rectifying strategy according to a depth confidence level, determining deviation rectifying quantity corresponding to the deviation rectifying strategy under a predefined safety space geometric constraint, generating a transition track based on the deviation rectifying quantity, and controlling the surgical instrument to execute the transition track. The application can realize the intelligent decision closed loop with the operation step semantics as the core and the multi-mode perception fusion.

Inventors

LI YAO
XIE YING
CHEN HAO
GONG JUNJIE
WANG PENG
Fan Peihan

Assignees

成都博恩思医学机器人有限公司

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. An intelligent control method of a microsurgical robot is characterized by comprising the following steps: When each surgical step is executed based on a preset step time sequence library, synchronously acquiring visual information and multidimensional force information aiming at a surgical instrument through a binocular endoscope and a wrist force sensor; Calculating a visual deviation of an instrument tip based on the visual information, and calculating a force abnormality index based on the multi-dimensional force information; taking the currently executed operation step as a decision context, and judging whether a preset abnormal triggering condition is met or not based on the visual deviation and the force abnormality index; If yes, suspending the currently executed operation step, determining a corresponding deviation rectifying strategy according to the depth confidence level of the binocular endoscope, and determining deviation rectifying quantity which corresponds to the deviation rectifying strategy and is used for correcting the pose or operation parameters of the instrument under the predefined safety space geometric constraint; And generating a transition track based on the deviation rectifying quantity, and controlling the surgical instrument to execute the transition track.
2. The intelligent control method of a microsurgical robot according to claim 1, wherein determining a corresponding deviation rectifying strategy according to the depth confidence level of the binocular endoscope comprises: When the depth confidence level is high, determining that the corresponding deviation correcting strategy is to correct the full degree of freedom of the instrument pose based on the three-dimensional space pose deviation; when the depth confidence level is the middle, determining a corresponding deviation rectifying strategy to limit the deviation rectifying amplitude of the three-dimensional space, and carrying out mixed deviation rectifying by combining the pixel deviation of the two-dimensional image; and when the deep confidence level is low, determining that the corresponding deviation rectifying strategy is triggering safety maintenance or requesting manual intervention.
3. The intelligent control method of a microsurgical robot according to claim 1, wherein before the step of determining whether a preset abnormal trigger condition is satisfied based on the visual deviation and the force abnormality index in the decision context, the method comprises: performing double-target fixation and epipolar correction on the binocular endoscope, and limiting a parallax searching range of stereo matching in a parallax interval corresponding to a preset working distance range of the surgical robot; Performing stereo matching in the area of the parallax interval to obtain parallax data; calculating depth information of the instrument tip according to the parallax data; And calculating to obtain the depth confidence based on at least one of left-right consistency, uniqueness of a matching result, texture characteristics of an image area and time consistency of depth information in the stereo matching process.
4. The intelligent control method of a microsurgical robot of claim 1, further comprising: Identifying and tracking pixel coordinates of an instrument tip in an imaging image of the binocular endoscope; calculating the pixel offset between the pixel coordinates and the center of the image field of view; and if the pixel offset exceeds a preset offset value, controlling the binocular endoscope to move according to a planned path so as to enable the instrument tip to be centered.
5. The intelligent control method of a microsurgical robot according to claim 4, wherein the controlling the binocular endoscope to move according to a planned path comprises: judging whether the preset collision conflict or the preset shielding conflict is triggered if the binocular endoscope executes the in-return control; if not, controlling the binocular endoscope to move based on the planned path controlled in the return process; If so, solving an optimal viewing pose of the binocular endoscope under the condition of collision avoidance constraint and predefined safety space geometric constraint by maximizing the visibility of the instrument tip and/or maximizing the image coverage of a key area as an optimization target, and controlling the binocular endoscope to move to the optimal viewing pose.
6. The intelligent control method of a microsurgical robot according to claim 1, wherein the step sequence library includes a step sequence, an action primitive, a transition condition, and a rollback policy, and before synchronously acquiring visual information and multidimensional force information for a surgical instrument through a binocular endoscope and a wrist force sensor when each surgical step is performed based on the preset step sequence library, further comprising: constructing a standardized sequence of steps for each surgical subtask in a target surgical scene, the surgical subtask including at least one of threading, suturing, and knotting; Associating at least one action primitive for each surgical step in the sequence of steps, the action primitive comprising at least one of a pose trajectory primitive for controlling the surgical instrument, a grip primitive, a force primitive, and a lens follower primitive for controlling the binocular endoscope; setting a corresponding transition condition for each surgical step, the transition condition including at least one of an instrument in-place threshold, a visual alignment threshold, a force threshold, a time threshold, and a confidence threshold; A corresponding retraction strategy is preset for each surgical step, including at least one of reclamping, retracting to a safe pose, re-performing an alignment action, and switching to a manual control mode.
7. The intelligent control method of a microsurgical robot of claim 1, further comprising: Responding to a manual takeover instruction, freezing an automatic control output based on the step time sequence library, and transferring control rights of the surgical instrument and the binocular endoscope to manual operation; After the manual operation is finished, collecting current system state vectors, wherein the system state vectors comprise instrument pose, clamping state, target point visibility, force/moment state and task execution confidence; And matching the system state vector with the expected states of all operation steps in the step time sequence library, determining that the operation step with the highest matching degree is a resynchronization node, and executing the next operation step of the resynchronization node.
8. The intelligent control method of a microsurgical robot according to claim 7, wherein the expected states include a reference state vector and a tolerance state vector, the matching the system state vector with the expected states of each surgical step in the step sequence library, determining that the surgical step with the highest matching degree is a resynchronisation node, includes: taking an operation step, in which a state range formed by a reference state vector and a tolerance state vector is overlapped with each component of the system state vector, as a candidate step node; Calculating normalized deviation of the system state vector and the reference state vector of each candidate step node in each state dimension; And calculating the matching score of each candidate step node based on the normalized deviation, and determining the candidate step node with the highest matching score as the resynchronization node.
9. An intelligent control system for a microsurgical robot, comprising: A sensing module comprising a binocular endoscope and a wrist force sensor; The execution module comprises a robot execution mechanism and a surgical instrument; Decision and control module comprising a control calculation unit for performing the intelligent control method of a microsurgical robot as claimed in any of claims 1 to 8.
10. A computer storage medium storing executable instructions that when executed by a processor cause the processor to perform the intelligent control method of a microsurgical robot in accordance with any one of claims 1 to 8.

Description

Intelligent control method and system for microsurgery robot Technical Field The application relates to the technical field of medical robot control, in particular to an intelligent control method and system of a microsurgical robot. Background With the expansion of the robot-assisted surgery technology to finer intra-cavity, oral and open surgery fields, the demand for robots to automatically perform programmed fine operations such as suturing, knotting and the like is becoming urgent. However, current surgical robotic systems, when dealing with such high-continuity demanding tasks, generally expose a core contradiction in that the systems lack the intelligence to uniformly understand and deal with preset, rigid programmed operating steps and dynamic, uncertain, complex constraint-filled physical surgical environments. When common dynamic interferences such as tissue traction displacement, instrument slipping, visual field loss and the like occur in the operation, most of the existing systems only can adopt simple scram or trigger fixed alarm, so that the operation flow is forced to be interrupted, all treatment responsibilities are returned to doctors, and the continuity and the overall efficiency of automatic execution are severely restricted. Therefore, how to provide the robot system with the capability of actively and smoothly assisting in adjusting and safely continuing the operation under the dynamic interference is a key bottleneck for improving the autonomy level of the surgical robot. The prior art has made various attempts to improve the coping capability of the system, mainly focusing on two aspects, namely, the capability of enhancing a single sensing dimension, such as developing a more robust visual tracking algorithm to improve positioning accuracy or adopting a more sensitive force sensor to detect contact force abnormality earlier, and the planning and constraint mechanism of the enhanced system, such as designing a more complex obstacle avoidance path or setting a more strict virtual safety boundary. However, these improvements still have an uncomplicated fault at all, a lack of an intelligent decision layer based on the current surgical context and the perceived reliability of itself, between the perceived abnormal signal and the corrective action that should be performed. This results in the system being fractured and passive in response logic in the face of anomalies-either unresponsive due to conservative threshold settings or frequent interruptions in flow due to false triggers, and eventually often falling into an inefficient loop of "perceptual anomalies, emergency stops, manual resets, restarts", failing to form an intelligent closed loop of "perceptual-evaluation-decision-adaptive splice". It follows that a further disadvantage of the prior art solution is the mechanical and discontinuous nature of its response pattern. Therefore, it is currently needed to propose a new system control paradigm that can implement multi-modal sensing fusion intelligent decision-making closed loop with surgical step semantics as a core. Disclosure of Invention The application aims to provide an intelligent control method and system for a microsurgical robot, so as to solve the problems. To achieve the above object, in a first aspect, the present application provides an intelligent control method of a microsurgical robot, the method comprising: When each surgical step is executed based on a preset step time sequence library, synchronously acquiring visual information and multidimensional force information aiming at a surgical instrument through a binocular endoscope and a wrist force sensor; Calculating a visual deviation of an instrument tip based on the visual information, and calculating a force abnormality index based on the multi-dimensional force information; taking the currently executed operation step as a decision context, and judging whether a preset abnormal triggering condition is met or not based on the visual deviation and the force abnormality index; If yes, suspending the currently executed operation step, determining a corresponding deviation rectifying strategy according to the depth confidence level of the binocular endoscope, and determining deviation rectifying quantity which corresponds to the deviation rectifying strategy and is used for correcting the pose or operation parameters of the instrument under the predefined safety space geometric constraint; And generating a transition track based on the deviation rectifying quantity, and controlling the surgical instrument to execute the transition track. In some embodiments, the determining a corresponding deskew strategy based on the depth confidence level of the binocular endoscope comprises: When the depth confidence level is high, determining that the corresponding deviation correcting strategy is to correct the full degree of freedom of the instrument pose based on the three-dimensional space pose deviation; when the depth c