Visual positioning and picking method and device for movable robot blood gas consumables

Abstract

The application provides a visual positioning, taking and placing method and device for a movable robot blood gas consumable, relates to the technical field of medical automation, and comprises the following steps: acquiring an image through a visual sensor, extracting blood gas consumable features to establish a pose model, calculating adaptive clamp parameters, generating a collision-free trajectory, executing a force control strategy based on material characteristics, adjusting the grabbing strength in real time, and finally accurately placing the blood gas consumable in a target area. The application improves the accuracy and stability of blood gas consumable taking and placing, reduces the risk of damage, adapts to different specifications of consumables, and significantly improves the efficiency of medical automation operations.

Description

Technical Field

[0001] This invention relates to the field of medical automation technology, and in particular to a visual positioning and placement method and device for mobile robotic blood gas consumables. Background Technology

[0002] Blood gas analysis is a crucial component of clinical laboratory testing, used to assess a patient's respiratory and metabolic status. In medical laboratory environments, the handling of blood gas consumables requires precise, aseptic, and efficient procedures. Traditional management of blood gas consumables relies primarily on manual operation, which is not only inefficient but also prone to contamination and sample misinterpretation. With the development of medical automation technology, the application of robot-assisted systems in the pretreatment stage of blood gas analysis is gaining increasing attention.

[0003] Traditional blood gas handling systems lack effective visual perception and pose estimation capabilities, failing to accurately identify the spatial position and orientation of different types of blood gas consumables. This results in low grasping accuracy and a high risk of missed or incorrect grasping. Existing systems typically employ fixed clamp structures, which are ill-suited for blood gas consumables of various sizes and shapes. They are particularly lacking in effective force control strategies for fragile or easily deformable consumables made of special materials, easily leading to sample damage during operation. Furthermore, traditional systems lack intelligent path planning and collision avoidance mechanisms during the transfer of blood gas consumables from the storage area to the placement area. This makes them unable to cope with complex and changing laboratory environments, posing safety hazards and resulting in low operational efficiency. Summary of the Invention

[0004] This invention provides a visual positioning and placement method and apparatus for mobile robot blood gas consumables, which can solve the problems in the prior art.

[0005] A first aspect of the present invention provides a visual positioning and placement method for blood gas analyzers used in a mobile robot, comprising:

[0006] The robot's vision sensors acquire visual images of blood gas consumables, storage areas, and placement areas.

[0007] The geometric contour features of the blood gas consumables are extracted from the visual image, and a consumable pose model is constructed by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates and pose angle of the blood gas consumables.

[0008] Based on the size parameters and the consumable type identifier, the clamping distance and contact surface shape of the robot's adaptive gripper are calculated, and the gripper configuration parameters are generated.

[0009] Based on the three-dimensional spatial coordinates, the position of the blood gas consumable is spatially correlated with the robot's current pose to generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable.

[0010] The robot is driven to move to the storage area according to the collision-free motion trajectory, the approach direction is determined according to the gripper configuration parameters and the posture angle, and a force control strategy is generated based on the material characteristics of the blood gas consumable; during the grasping process, contact force feedback information is acquired in real time, and the force control strategy is dynamically corrected using the contact force feedback information;

[0011] The spatial coordinates and placement posture constraint information of the placement area are extracted from the visual image, and the corresponding placement position is matched according to the consumable type identifier. A placement trajectory to the placement position is generated, and the placement action is performed according to the placement trajectory.

[0012] The geometric contour features of the blood gas consumable are extracted from the visual image, and a consumable pose model is constructed by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates, and pose angle of the blood gas consumable, including:

[0013] The pixel-level edge contour of the blood gas consumable in the visual image is extracted by an edge detection operator, and the pixel-level edge contour is fitted to obtain the two-dimensional geometric contour features of the blood gas consumable; based on the two-dimensional geometric contour features, feature matching is performed to determine the consumable type identifier and size parameters of the blood gas consumable.

[0014] The spatial coordinate system transformation relationship from the image coordinate system to the world coordinate system is established by using the intrinsic and extrinsic parameter matrices of the vision sensor.

[0015] The pixel coordinates of feature points in the two-dimensional geometric contour features are extracted. Based on the spatial coordinate system transformation relationship, the pixel coordinates are converted into normalized coordinates in the camera coordinate system using the intrinsic parameter matrix. Then, the normalized coordinates are converted into the three-dimensional spatial coordinates in the world coordinate system using the extrinsic parameter matrix.

[0016] Calculate the direction vector of the major axis direction in the two-dimensional geometric contour feature in the image coordinate system, and transform it to the world coordinate system through the spatial coordinate system transformation relationship. Based on the direction vector in the world coordinate system, calculate the rotation angle of the blood gas consumable about each coordinate axis of the world coordinate system to obtain the attitude angle of the blood gas consumable relative to the world coordinate system.

[0017] Based on the size parameters and the consumable type identifier, the clamping distance and contact surface shape of the robot's adaptive gripper are calculated, and the gripper configuration parameters are generated, including:

[0018] A three-dimensional geometric mesh for the blood gas consumable is constructed based on the size parameters, and the yield strength of the blood gas consumable is obtained according to the consumable type identifier.

[0019] Multiple sets of candidate clamping configurations are set, and clamping force boundary conditions corresponding to each set of candidate clamping configurations are applied to the three-dimensional geometric mesh;

[0020] Finite element analysis was performed on the three-dimensional geometric mesh after applying the clamping force boundary condition to obtain the stress distribution field and deformation distribution field of the blood gas consumable under each set of candidate clamping configurations.

[0021] Extract the maximum stress value from the stress distribution field, extract the maximum deformation value from the deformation distribution field, and select the target clamping configuration from multiple candidate clamping configurations that makes the maximum stress value lower than the yield strength and the maximum deformation value the smallest.

[0022] Extract the target clamping distance value and the target contact surface curvature value from the target clamping configuration, calculate the displacement control amount of the drive mechanism of the adaptive fixture based on the target clamping distance value, calculate the curvature control amount of the shape adjustment mechanism of the adaptive fixture based on the target contact surface curvature value, and encapsulate the displacement control amount and the curvature control amount into the fixture configuration parameters.

[0023] Based on the three-dimensional spatial coordinates, the position of the blood gas consumable is spatially correlated with the robot's current pose to generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable, including:

[0024] The robot's current position coordinates are used as the starting coordinates of the motion trajectory, and the three-dimensional spatial coordinates are used as the ending coordinates of the motion trajectory. A spatial vector is constructed in the world coordinate system from the starting coordinates to the ending coordinates. The length of the spatial vector is calculated as the motion path length. Intermediate path nodes are set between the starting coordinates and the ending coordinates along the direction of the spatial vector. The number of intermediate path nodes is determined according to the motion path length.

[0025] In the world coordinate system, a three-dimensional envelope volume is established to represent the space occupied by the obstacle. Each intermediate path node is checked in turn to determine whether it is located inside or on the boundary of the three-dimensional envelope volume. If so, the current intermediate path node is marked as a collision node. The shortest distance vector between the collision node and the surface of the three-dimensional envelope volume is calculated. The collision node is offset in the opposite direction of the shortest distance vector to obtain the intermediate path node sequence after obstacle avoidance.

[0026] By connecting the starting point coordinates, the sequence of intermediate path nodes after obstacle avoidance, and the ending point coordinates in chronological order, the collision-free motion trajectory is generated.

[0027] Driving the robot to move to the storage area according to the collision-free motion trajectory, determining the approach direction based on the gripper configuration parameters and the attitude angle, and generating a force control strategy based on the material properties of the blood gas consumables includes:

[0028] The collision-free motion trajectory is converted into an angle command sequence for each joint of the robot, and the angle commands of each joint in the angle command sequence are executed sequentially in chronological order to drive the robot to move to the storage area;

[0029] Based on the fixture configuration parameters, the contact normal vector when the adaptive fixture contacts the blood gas consumable is calculated according to the target clamping distance value and the target contact surface curvature value, and the direction of the contact normal vector in the world coordinate system is determined as the approach direction.

[0030] Obtain the corresponding material hardness value and compressive strength value according to the consumable type identifier, calculate the maximum contact force threshold that can be applied during the gripping process based on the material hardness value, and calculate the maximum deformation threshold that can be generated during the gripping process based on the compressive strength value;

[0031] The maximum contact force threshold is calculated based on the material hardness value, and a contact force control curve is generated. The maximum deformation threshold is calculated based on the compressive strength value, and a deformation monitoring threshold is set. The force control strategy is set according to the approach direction, the contact force control curve, and the deformation monitoring threshold.

[0032] Real-time acquisition of contact force feedback information during the grasping process, and dynamic correction of the force control strategy using the contact force feedback information, including:

[0033] During the gripping action of the adaptive clamp, the normal contact force value and tangential contact force value on the contact surface between the adaptive clamp and the blood gas consumable are collected in real time as the contact force feedback information.

[0034] Extract the target contact force value corresponding to the current moment from the force control strategy, calculate the force deviation value between the normal contact force value and the target contact force value, and determine that a force correction operation needs to be performed when the force deviation value exceeds a preset force deviation threshold or the tangential contact force value exceeds the preset tangential force threshold.

[0035] When performing force correction, the absolute value of the force deviation is taken as the force correction amplitude. The positive and negative directions of the force deviation are combined with the force correction amplitude to calculate the force correction amount and the corresponding direction of action. Based on the force correction amount, the clamping force output of the adaptive gripper is adjusted until the actual contact force value applied by the adaptive gripper converges to the target contact force value, thus obtaining the adjusted robot force control parameters. The adjusted robot force control parameters are then updated in the force control strategy.

[0036] The spatial coordinates and placement posture constraint information of the placement area are extracted from the visual image, and the corresponding placement position is matched according to the consumable type identifier to generate a placement trajectory to the placement position, including:

[0037] The visual image is segmented into regions to identify the coordinates of the center point of the placement area, and the three-dimensional coordinates of the center point are used as the spatial position coordinates.

[0038] The placement direction and spacing of other blood gas consumables already placed in the placement area are identified from the visual image. The posture angle constraint range of the placement area is determined based on the placement direction, and the position offset constraint range of the placement area is determined based on the spacing, thereby obtaining the placement posture constraint information.

[0039] Based on the consumable type identifier, the priority placement order and preset placement slots of different types of blood gas consumables in the placement area are set to obtain placement rules; based on the placement rules, the spatial position coordinates and the placement posture constraint information, the corresponding placement position is determined;

[0040] The robot calculates the spatial displacement vector and attitude rotation vector between its current pose and the placement position. Based on the spatial displacement vector and attitude rotation vector, it plans a motion path to the placement position and sets corresponding timestamps and velocity parameters to generate the placement trajectory containing time and velocity information.

[0041] A second aspect of the present invention provides a visual positioning and placement device for mobile robotic blood gas consumables, comprising:

[0042] The first unit is used to acquire visual images of blood gas consumables, storage areas, and placement areas through the robot's vision sensors;

[0043] The second unit is used to extract the geometric contour features of the blood gas consumable from the visual image, and construct the consumable pose model by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates and pose angle of the blood gas consumable.

[0044] The third unit is used to calculate the clamping distance and contact surface shape of the robot's adaptive gripper based on the size parameters and the consumable type identifier, and generate gripper configuration parameters.

[0045] The fourth unit is used to spatially associate the position of the blood gas consumable with the robot's current pose based on the three-dimensional spatial coordinates, and generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable.

[0046] The fifth unit is used to drive the robot to move to the storage area according to the collision-free motion trajectory, determine the approach direction based on the gripper configuration parameters and the posture angle, and generate a force control strategy based on the material characteristics of the blood gas consumable; during the grasping process, contact force feedback information is acquired in real time, and the force control strategy is dynamically corrected using the contact force feedback information;

[0047] The sixth unit is used to extract the spatial coordinates and placement posture constraint information of the placement area from the visual image, match the corresponding placement position according to the consumable type identifier, generate a placement trajectory to the placement position, and perform the placement action according to the placement trajectory.

[0048] A third aspect of the embodiments of the present invention,

[0049] An electronic device is provided, comprising:

[0050] processor;

[0051] Memory used to store processor-executable instructions;

[0052] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0053] Fourth aspect of the present invention,

[0054] A computer-readable storage medium is provided, having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0055] The beneficial effects of this application are as follows:

[0056] By extracting the geometric contour features of blood gas analyzers using machine vision and constructing a pose model, accurate identification and spatial positioning of different types of blood gas analyzers are achieved, significantly improving the adaptability and accuracy of blood gas analyzer handling. The fixture configuration is adaptively adjusted based on the analyzer type identification and size parameters, resolving compatibility issues for handling different specifications of blood gas analyzers and improving versatility. A force control strategy is generated based on the material characteristics of the blood gas analyzers and dynamically corrected through real-time contact force feedback, effectively preventing damage to fragile and deformable blood gas analyzers and improving the safety and reliability of the gripping process. Attached Figure Description

[0057] Figure 1 This is a flowchart illustrating the visual positioning and placement method for blood gas consumables in a mobile robot according to an embodiment of the present invention.

[0058] Figure 2 A schematic diagram of the calculation process for optimizing adaptive fixture configuration parameters. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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.

[0060] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0061] Figure 1 This is a flowchart illustrating the visual positioning and placement method for blood gas analyzers in a mobile robot according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0062] The robot's vision sensors acquire visual images of blood gas consumables, storage areas, and placement areas.

[0063] The geometric contour features of the blood gas consumables are extracted from the visual image, and a consumable pose model is constructed by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates and pose angle of the blood gas consumables.

[0064] Based on the size parameters and the consumable type identifier, the clamping distance and contact surface shape of the robot's adaptive gripper are calculated, and the gripper configuration parameters are generated.

[0065] Based on the three-dimensional spatial coordinates, the position of the blood gas consumable is spatially correlated with the robot's current pose to generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable.

[0066] The robot is driven to move to the storage area according to the collision-free motion trajectory, the approach direction is determined according to the gripper configuration parameters and the posture angle, and a force control strategy is generated based on the material characteristics of the blood gas consumable; during the grasping process, contact force feedback information is acquired in real time, and the force control strategy is dynamically corrected using the contact force feedback information;

[0067] The spatial coordinates and placement posture constraint information of the placement area are extracted from the visual image, and the corresponding placement position is matched according to the consumable type identifier. A placement trajectory to the placement position is generated, and the placement action is performed according to the placement trajectory.

[0068] In one optional implementation, the geometric contour features of the blood gas consumable are extracted from the visual image, and a consumable pose model is constructed by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates, and pose angle of the blood gas consumable, including:

[0069] The pixel-level edge contour of the blood gas consumable in the visual image is extracted by an edge detection operator, and the pixel-level edge contour is fitted to obtain the two-dimensional geometric contour features of the blood gas consumable; based on the two-dimensional geometric contour features, feature matching is performed to determine the consumable type identifier and size parameters of the blood gas consumable.

[0070] The spatial coordinate system transformation relationship from the image coordinate system to the world coordinate system is established by using the intrinsic and extrinsic parameter matrices of the vision sensor.

[0071] The pixel coordinates of feature points in the two-dimensional geometric contour features are extracted. Based on the spatial coordinate system transformation relationship, the pixel coordinates are converted into normalized coordinates in the camera coordinate system using the intrinsic parameter matrix. Then, the normalized coordinates are converted into the three-dimensional spatial coordinates in the world coordinate system using the extrinsic parameter matrix.

[0072] Calculate the direction vector of the major axis direction in the two-dimensional geometric contour feature in the image coordinate system, and transform it to the world coordinate system through the spatial coordinate system transformation relationship. Based on the direction vector in the world coordinate system, calculate the rotation angle of the blood gas consumable about each coordinate axis of the world coordinate system to obtain the attitude angle of the blood gas consumable relative to the world coordinate system.

[0073] The robot's vision sensors acquire visual images of the blood gas consumables, storage area, and placement area. Stable lighting conditions are ensured during image acquisition to avoid strong glare or shadows. The acquired visual images are stored in color or grayscale format with a resolution of at least 1280×960 pixels to guarantee the accuracy of subsequent geometric feature extraction.

[0074] The visual image is preprocessed, including Gaussian filtering for noise reduction and grayscale conversion. Then, the Canny edge detection operator is used to extract the edges of the consumables. The Canny operator performs smoothing through Gaussian filtering, calculates the image gradient magnitude and direction, performs non-maximum suppression, and determines the edges through dual threshold detection. For consumables such as blood gas syringes, setting a low threshold of 50 and a high threshold of 150 can effectively separate the foreground from the background.

[0075] The acquired pixel-level edges are fitted with contours to transform the discrete set of edge points into a mathematically described geometric shape. For approximately cylindrical consumables such as blood gas syringes, an ellipse fitting algorithm is used. Specifically, the least squares method is applied to fit the ellipse equation to obtain the center coordinates (x0, y0), the major semi-axis a, the minor semi-axis b, and the rotation angle θ. These parameters constitute the two-dimensional geometric contour features of the blood gas consumable.

[0076] Based on the extracted two-dimensional geometric contour features, feature matching is performed to determine the type and size parameters of blood gas consumables. A consumable feature library is pre-established, containing standard geometric parameters for each type of consumable. The similarity between the feature vector of the current consumable and each template in the feature library is calculated according to the type and size parameters, and the match with the highest similarity is selected as the recognition result. For example, if the ratio of the major axis to the minor axis of the extracted consumable is 10:1 and the length is approximately 100 mm, it matches as a standard blood gas injector; if the ratio is close to 1:1 and the diameter is approximately 13 mm, it matches as a blood gas collection tube.

[0077] Obtain the calibration parameters of the vision sensor. The intrinsic parameter matrix K of the vision sensor contains the focal length (fx, fy) and principal point coordinates (cx, cy) information, and the extrinsic parameter matrix [R|t] describes the rotation matrix R and translation vector t of the camera relative to the world coordinate system.

[0078] The pixel coordinates (u, v) of feature points in the 2D geometric contour are extracted, including key locations such as the center point and endpoints of the consumable. Based on the spatial coordinate system transformation relationship, the pixel coordinates are converted into normalized coordinates (xn, yn) in the camera coordinate system using an intrinsic parameter matrix. The transformation calculation is: xn = (u - cx) / fx, yn = (v - cy) / fy. For a blood gas injector, its two endpoints and center point can be extracted as feature points to obtain its normalized coordinates.

[0079] Combining known consumable size information and depth estimation algorithm, for a standard blood gas injector, whose diameter is known to be 10 mm, it can be used as a reference for depth recovery. By using the known physical size and image size ratio, the depth value z of the feature point is calculated. Using the normalized coordinates and depth value, the three-dimensional coordinates (xc, yc, zc) of the feature point in the camera coordinate system are calculated as follows: xc=xn×z, yc=yn×z, zc=z.

[0080] The 3D coordinates in the camera coordinate system are transformed to the world coordinate system using an extrinsic parameter matrix, calculated as: [Xw, Yw, Zw, 1] T =[R|t] -1 ×[xc, yc, zc, 1] T , where [Xw, Yw, Zw] are the three-dimensional spatial coordinates of the feature point in the world coordinate system.

[0081] Determine the major axis direction vector of the consumable in the image. For a blood gas injector, the major axis direction is the extension direction of the injector body. By extracting the rotation angle θ in the two-dimensional geometric contour features, the major axis direction vector (cos(θ), sin(θ)) in the image coordinate system can be obtained.

[0082] The direction vector in the image coordinate system is transformed to the camera coordinate system, and then to the world coordinate system through an extrinsic parameter matrix, yielding the direction vector Vw of the consumable in the world coordinate system. Based on Vw, the arctangent function is used to calculate the rotation angles of the consumable around the X, Y, and Z axes. This completes the construction of the blood gas consumable pose model, including consumable type identification, size parameters, three-dimensional spatial coordinates, and pose angles, providing accurate pose information for subsequent robotic arm grasping or other operations.

[0083] In one optional implementation, the clamping spacing and contact surface morphology of the robot's adaptive gripper are calculated based on the size parameters and the consumable type identifier, and the gripper configuration parameters are generated, including:

[0084] A three-dimensional geometric mesh for the blood gas consumable is constructed based on the size parameters, and the yield strength of the blood gas consumable is obtained according to the consumable type identifier.

[0085] Multiple sets of candidate clamping configurations are set, and clamping force boundary conditions corresponding to each set of candidate clamping configurations are applied to the three-dimensional geometric mesh;

[0086] Finite element analysis was performed on the three-dimensional geometric mesh after applying the clamping force boundary condition to obtain the stress distribution field and deformation distribution field of the blood gas consumable under each set of candidate clamping configurations.

[0087] Extract the maximum stress value from the stress distribution field, extract the maximum deformation value from the deformation distribution field, and select the target clamping configuration from multiple candidate clamping configurations that makes the maximum stress value lower than the yield strength and the maximum deformation value the smallest.

[0088] Extract the target clamping distance value and the target contact surface curvature value from the target clamping configuration, calculate the displacement control amount of the drive mechanism of the adaptive fixture based on the target clamping distance value, calculate the curvature control amount of the shape adjustment mechanism of the adaptive fixture based on the target contact surface curvature value, and encapsulate the displacement control amount and the curvature control amount into the fixture configuration parameters.

[0089] like Figure 2 As shown, the method includes:

[0090] The geometric dimensional parameters of blood gas consumables include key parameters such as length, diameter, and wall thickness. Using these parameters, a three-dimensional geometric mesh model of the blood gas consumable is constructed using finite element method (FEM). For example, a common blood gas injector can be simplified into a composite structure consisting of a cylinder and a conical head. A hexahedral mesh is used for the cylindrical part, while a tetrahedral mesh is used for more complex geometric parts, such as the conical joint. The mesh density is dynamically adjusted according to stress concentration areas. Typically, the mesh size is set below 0.1 mm in areas of expected stress concentration, while it can be appropriately increased to 0.5 mm in non-critical areas to balance computational accuracy and efficiency.

[0091] By querying a pre-defined database of blood gas consumable materials, the corresponding material yield strength is obtained based on the consumable type identifier. For example, the yield strength of a common polypropylene blood gas injector is approximately 30 MPa, while that of polycarbonate consumables can reach 60 MPa. This material property data will serve as an important basis for subsequent assessments of clamping safety.

[0092] Multiple candidate clamping configurations are defined, which can include different combinations of parameters such as clamping position, clamping contact area, contact surface shape, and clamping force. For example, five different clamping spacings can be designed, ranging from 80% to 120% of the consumable diameter, varying in 5% increments; simultaneously, three different contact surface curvatures can be designed: planar contact, arc contact (curvature radius 1.5 times the consumable diameter), and conformal contact (curvature radius matching the consumable diameter). For each configuration combination, the applied clamping force starts at 5N and increases to 15N in 1N increments, forming a complete candidate clamping configuration parameter space.

[0093] Considering the contact characteristics between the fixture and the consumable, including the coefficient of friction (usually 0.2-0.4) and contact stiffness, apply corresponding force boundary conditions to the mesh model to ensure that the boundary conditions can accurately reflect the actual clamping scenario.

[0094] A nonlinear solver is used to perform finite element analysis on the three-dimensional geometric mesh after applying the clamping force boundary condition. The elastic-plastic properties and large deformation effects of the material are considered. During the solution process, mesh independence is verified to ensure that the mesh generation accuracy does not affect the calculation results. An incremental loading strategy is adopted to apply the clamping force step by step to capture the nonlinear response.

[0095] Finite element analysis can yield detailed stress distribution fields for each clamping configuration, including von Mises equivalent stress and principal stresses; it also provides complete information on deformation distribution fields, including displacement and strain fields. These data form the basis for subsequent clamping configuration optimization.

[0096] The analysis results for each clamping configuration were post-processed to extract key indicators: the maximum von Mises equivalent stress value and its location, the maximum displacement and its distribution characteristics. The maximum stress value was compared with the material yield strength, and a safety factor of 1.5 was set, requiring the maximum stress value to not exceed 66.7% of the yield strength. Among all configurations that met the safety stress condition, the configuration that minimized the maximum deformation was further selected as the target clamping configuration.

[0097] For the target clamping configuration, extract its key parameters: target clamping distance value and target contact surface curvature value. Considering the mechanical structure characteristics of the adaptive fixture, convert the clamping distance value into the rotation angle of the drive motor or the displacement of the linear actuator. At the same time, convert the target contact surface curvature value into the control parameters of the shape adjustment mechanism, such as the pneumatic control pressure of the flexible contact surface or the deformation of the mechanical structure.

[0098] Using the above method, the optimal clamping parameters can be automatically calculated based on the characteristics of different blood gas consumables, ensuring that the clamping process is both safe and reliable, and will not damage the blood gas consumables, thereby improving the accuracy and safety of robot operation.

[0099] In one optional implementation, based on the three-dimensional spatial coordinates, the location of the blood gas consumable is spatially correlated with the robot's current pose to generate a collision-free motion trajectory from the current pose to the location of the blood gas consumable, including:

[0100] The robot's current position coordinates are used as the starting coordinates of the motion trajectory, and the three-dimensional spatial coordinates are used as the ending coordinates of the motion trajectory. A spatial vector is constructed in the world coordinate system from the starting coordinates to the ending coordinates. The length of the spatial vector is calculated as the motion path length. Intermediate path nodes are set between the starting coordinates and the ending coordinates along the direction of the spatial vector. The number of intermediate path nodes is determined according to the motion path length.

[0101] In the world coordinate system, a three-dimensional envelope volume is established to represent the space occupied by the obstacle. Each intermediate path node is checked in turn to determine whether it is located inside or on the boundary of the three-dimensional envelope volume. If so, the current intermediate path node is marked as a collision node. The shortest distance vector between the collision node and the surface of the three-dimensional envelope volume is calculated. The collision node is offset in the opposite direction of the shortest distance vector to obtain the intermediate path node sequence after obstacle avoidance.

[0102] By connecting the starting point coordinates, the sequence of intermediate path nodes after obstacle avoidance, and the ending point coordinates in chronological order, the collision-free motion trajectory is generated.

[0103] The robot's current position coordinates are obtained as the starting coordinates of the motion trajectory, and the detected three-dimensional spatial coordinates of the blood gas consumables are used as the ending coordinates of the motion trajectory. A spatial vector from the starting coordinates to the ending coordinates is constructed in the world coordinate system. This vector can be represented as the difference vector between the ending coordinates and the starting coordinates. The Euclidean distance of this spatial vector is calculated as the motion path length.

[0104] The number of intermediate path nodes is determined based on the calculated path length. For example, the number of nodes can be determined by setting an intermediate path node at fixed intervals (such as 10 centimeters). In practical applications, if the path length is one meter and a node is placed every 10 centimeters, a total of nine intermediate path nodes are needed. These nodes are evenly distributed along the spatial vector direction between the start and end points to form the initial path plan.

[0105] For fixed obstacles (such as workbenches, instruments, etc.) in laboratory or medical settings, their geometric information can be obtained through 3D modeling or depth camera scanning. For each obstacle, a corresponding 3D envelope, such as a cube, sphere, or cylinder, is constructed to simplify collision detection calculations. These 3D envelopes have clearly defined position and size parameters in the world coordinate system.

[0106] Collision detection is performed on each intermediate path node. Geometric calculations are used to sequentially check whether each intermediate path node is located inside or on the boundary of any 3D envelope. For example, for a cube envelope, it is determined whether the three coordinate components of the node are all within the corresponding dimensions of the cube; for a sphere envelope, it is calculated whether the distance from the node to the center of the sphere is less than or equal to the radius of the sphere. When a node is found to be inside or on the boundary of an obstacle envelope, it is marked as a collision node.

[0107] For each marked collision node, calculate its shortest distance vector to the corresponding 3D envelope surface. For a cubic envelope, find the distance from the node to each surface and select the direction corresponding to the minimum value. For a spherical envelope, the shortest distance vector is the unit vector from the center of the sphere to the node multiplied by (the radius of the sphere minus the distance from the node to the center of the sphere). After obtaining the shortest distance vector, offset the collision node in the opposite direction of this vector. The offset distance can be set to the shortest distance plus a safety margin (e.g., five centimeters) to ensure that the adjusted node maintains a sufficient safe distance from the obstacle.

[0108] After obstacle avoidance, a series of intermediate path node sequences after obstacle avoidance are obtained. In order to make the motion smoother, spline curve fitting or Bézier curve interpolation can be performed on these nodes to generate a continuous and smooth path curve. The starting coordinates, the intermediate path node sequence after obstacle avoidance and the ending coordinates are connected in time order to form a complete collision-free motion trajectory.

[0109] In practical applications, such as when a medical robot needs to move from the nurse's station to a blood gas analyzer to retrieve or place blood gas consumables, this method can effectively avoid medical carts, hospital beds, or other moving obstacles along the way. When the robot detects that the blood gas consumable is located at the sample inlet of the blood gas analyzer, it obtains the robot's current coordinates at the nurse's station as the starting point and the location of the blood gas consumable as the ending point, and plans a safe path that avoids moving hospital beds and IV stands in the middle of the corridor, enabling the robot to successfully reach the target location to perform the retrieval or placement task.

[0110] When new obstacles appear in the environment or the positions of existing obstacles change, the robot can update obstacle information in real time and replan the remaining path based on the current robot position to ensure the safety of the movement process. This dynamic planning capability is particularly important in medical environments where there is frequent movement of people and equipment.

[0111] In one optional implementation, driving the robot to move to the storage area according to the collision-free motion trajectory, determining the approach direction based on the gripper configuration parameters and the attitude angle, and generating a force control strategy based on the material properties of the blood gas consumables includes:

[0112] The collision-free motion trajectory is converted into an angle command sequence for each joint of the robot, and the angle commands of each joint in the angle command sequence are executed sequentially in chronological order to drive the robot to move to the storage area;

[0113] Based on the fixture configuration parameters, the contact normal vector when the adaptive fixture contacts the blood gas consumable is calculated according to the target clamping distance value and the target contact surface curvature value, and the direction of the contact normal vector in the world coordinate system is determined as the approach direction.

[0114] Obtain the corresponding material hardness value and compressive strength value according to the consumable type identifier, calculate the maximum contact force threshold that can be applied during the gripping process based on the material hardness value, and calculate the maximum deformation threshold that can be generated during the gripping process based on the compressive strength value;

[0115] The maximum contact force threshold is calculated based on the material hardness value, and a contact force control curve is generated. The maximum deformation threshold is calculated based on the compressive strength value, and a deformation monitoring threshold is set. The force control strategy is set according to the approach direction, the contact force control curve, and the deformation monitoring threshold.

[0116] The collision-free trajectory points are converted into a sequence of angle values ​​in the robot's joint space. For example, for a six-DOF robot, each trajectory point corresponds to a set of joint angle values ​​(θ1, θ2, θ3, θ4, θ5, θ6). During the conversion process, the robot's workspace constraints, joint velocity constraints, and acceleration constraints are considered to ensure that the generated angle command sequence is within the robot's physical constraints.

[0117] The robot executes joint angle commands sequentially according to time order, employing a time interpolation algorithm to ensure smooth and continuous movement. At critical points, such as obstacle avoidance points, the robot is slowed down to ensure safety. By monitoring the error between the robot's actual and expected positions in real time, necessary trajectory fine-tuning is performed, enabling the robot to move precisely to the storage area where the blood gas consumables are located.

[0118] The location of the clamp contact point is determined based on the geometric characteristics of the blood gas consumable. For cylindrical consumables such as syringes, the contact point is usually located on its circumferential surface. A local geometric model is constructed based on the curvature value of the target contact surface. For areas with large curvature values ​​(such as the syringe wall), a local cylindrical surface model is established. A local coordinate system is established at each contact point, with the contact point as the origin and the normal vector pointing to the central axis of the cylindrical surface. The target clamping distance value is combined with the diameter parameter of the consumable to determine the expected positional relationship when the clamp is closed, and the ideal contact posture between the clamp and the consumable surface is calculated. Based on the direction of the normal vector in the local coordinate system of the contact point, combined with the overall posture of the consumable, the final contact normal vector is determined.

[0119] The contact normal vector in the local coordinate system is transformed to the world coordinate system through a coordinate transformation matrix. The approach direction is determined by taking into account the posture angle of the blood gas consumable, ensuring that the robot approaches the target object from the optimal angle, reducing the risk of collision and the probability of clamping failure.

[0120] The corresponding material hardness and compressive strength values ​​are obtained based on the consumable type identifier. These material parameters are stored in a preset material parameter database. For example, the hardness value of a glass test tube is 6.5 on the Mohs scale and the compressive strength value is 70 MPa; while the hardness value of a plastic container is 3.0 on the Mohs scale and the compressive strength value is 30 MPa.

[0121] The maximum allowable contact force threshold during the gripping process is calculated based on the material hardness value. The relationship between material hardness and safety factor is considered to ensure that the force applied by the clamp is strong enough to grip the object without causing damage. For example, for glass, which has high hardness, the maximum contact force threshold can be set to 15 Newtons; while for plastic, which has lower hardness, the threshold is set to 8 Newtons.

[0122] By using a materials mechanics model, the relationship between compressive strength and deformation is established to ensure that the blood and gas consumables will not suffer permanent deformation or structural damage during the gripping process. For example, for glass with a compressive strength of 70 MPa, the maximum deformation threshold can be set to 0.01 mm; while for plastic with lower compressive strength, the threshold is 0.1 mm.

[0123] The contact force control curve is generated based on the material hardness value. This curve describes the change of the force applied by the clamp from the initial contact to the complete clamping process. It usually adopts a piecewise function form. In the initial stage, the contact force is slowly increased to ensure stable contact. In the middle stage, the force is gradually increased until the ideal clamping force is reached. In the final stage, a constant force value is maintained to maintain a stable gripping state.

[0124] The deformation monitoring threshold is set based on the compressive strength value. This threshold is usually set to 80% of the maximum deformation threshold and is used as an early warning value. During the gripping process, the deformation of the blood gas consumable is monitored in real time. Once it approaches the early warning value, the clamping force will be adjusted to prevent excessive deformation.

[0125] Based on the approach direction, contact force control curve, and deformation monitoring threshold, a force control strategy is comprehensively set. This strategy includes the following elements: approach speed control parameters to ensure that the robot approaches the target object at an appropriate speed; force feedback control parameters to adjust the force applied by the gripper in real time; deformation monitoring parameters to monitor the deformation state of the blood gas consumables; and an abnormal state handling strategy to define the response measures when an abnormal situation is detected.

[0126] By implementing the above technical solutions, it is possible to accurately grasp blood gas consumables of different materials and shapes, ensuring the safety and reliability of the grasping process and effectively avoiding damage to fragile and easily deformable blood gas consumables.

[0127] In one optional implementation, acquiring contact force feedback information in real time during the grasping process and dynamically correcting the force control strategy using the contact force feedback information includes:

[0128] During the gripping action of the adaptive clamp, the normal contact force value and tangential contact force value on the contact surface between the adaptive clamp and the blood gas consumable are collected in real time as the contact force feedback information.

[0129] Extract the target contact force value corresponding to the current moment from the force control strategy, calculate the force deviation value between the normal contact force value and the target contact force value, and determine that a force correction operation needs to be performed when the force deviation value exceeds a preset force deviation threshold or the tangential contact force value exceeds the preset tangential force threshold.

[0130] When performing force correction, the absolute value of the force deviation is taken as the force correction amplitude. The positive and negative directions of the force deviation are combined with the force correction amplitude to calculate the force correction amount and the corresponding direction of action. Based on the force correction amount, the clamping force output of the adaptive gripper is adjusted until the actual contact force value applied by the adaptive gripper converges to the target contact force value, thus obtaining the adjusted robot force control parameters. The adjusted robot force control parameters are then updated in the force control strategy.

[0131] When the adaptive gripper begins to grasp the blood gas consumable, a force sensor array is installed on the contact surface between the gripper and the consumable to collect the normal and tangential contact force values ​​in real time. This sensor array consists of multiple miniature pressure sensing units, capable of accurately capturing force information at different contact points and forming a complete contact force distribution map. The acquisition frequency is set to 100Hz to ensure the capture of rapidly changing contact states. The collected normal force value reflects the degree of compression of the blood gas consumable by the gripper, while the tangential force value reflects the change in frictional force that causes the consumable to slide.

[0132] After the contact force information is collected, the target contact force value corresponding to the current moment is extracted from the pre-set force control strategy database. This target contact force value is an ideal force value pre-calculated based on the material properties, shape parameters, and grasping task requirements of the blood gas consumable. For example, for fragile glass capillary blood gas consumables, the target contact force value is set to 0.5N to 1.5N; while for plastic-cased blood gas consumables, the target contact force value is set to 1.5N to 3N.

[0133] Calculate the force deviation between the actual normal contact force value and the target contact force value, and set a preset force deviation threshold of ±0.3N and a preset tangential force threshold of 0.5N. When the absolute value of the calculated force deviation exceeds 0.3N, or the tangential contact force value exceeds 0.5N, the force correction operation procedure is triggered.

[0134] When performing force correction, the absolute value of the force deviation is taken as the force correction amplitude. If the actual force value is greater than the target force value (force deviation is positive), the clamping force needs to be reduced; if the actual force value is less than the target force value (force deviation is negative), the clamping force needs to be increased. The force correction amount is calculated using a proportional-integral-derivative (PID) control algorithm.

[0135] The force correction is calculated based on the force deviation value, the integral value of the force deviation, and the rate of change of the force deviation. The force deviation e(t) at the current time t is calculated as the difference between the actual normal contact force value and the target contact force value. Then, the integral term ∫e(t)dt of the force deviation is calculated, which is the cumulative sum of all force deviations from the start of contact to the current time. Next, the differential term de(t) / dt of the force deviation is calculated, which is the difference between the current force deviation and the force deviation at the previous time divided by the sampling time interval. Finally, these three terms are weighted and summed, resulting in the force correction u(t) = Kp×e(t) + Ki×∫e(t)dt + Kd×de(t) / dt, where the proportional coefficient Kp is set to 0.8, the integral coefficient Ki is set to 0.15, and the differential coefficient Kd is set to 0.05. These parameters have been experimentally optimized to ensure rapid response without excessive oscillation.

[0136] After calculating the force correction, the clamping force output is adjusted through the actuator of the adaptive clamp. For pneumatic clamps, precise force control is achieved by adjusting the opening of the air pressure valve; for electric clamps, it is achieved by adjusting the motor torque. The adjustment process employs a gradual control strategy to avoid gripping instability caused by sudden changes in force value.

[0137] During force adjustment, the change in contact force is continuously monitored. When the difference between the actual contact force value and the target contact force value is less than 0.1N, the force value is considered to have converged to the target force value, completing a single force correction. At this time, the current clamping position, force control parameters, and corresponding actuator state are recorded to form the adjusted robot force control parameter set.

[0138] The adjusted robot force control parameters are updated in the force control strategy database as a reference for subsequent similar grasping tasks. Simultaneously, a mapping relationship between the force control parameters and the characteristics of blood gas consumables is constructed, forming an adaptive learning mechanism. This mapping relationship is continuously optimized with the increase in the number of grasping attempts, enabling the robot to gradually master the optimal grasping parameters for different types of blood gas consumables.

[0139] In practical applications, when grasping a specific type of blood gas injector, the initial force control parameter is set to a normal force of 2.0N. During the grasping process, if the actual measured normal force reaches 2.5N, exceeding the preset deviation threshold of 0.3N, force correction is immediately triggered. The calculated force deviation is +0.5N, indicating a need to reduce the clamping force. The PID control algorithm calculates a force correction of -0.45N. The clamping actuator then reduces the clamping force until the normal force drops to 2.1N, with the difference from the target value less than 0.1N, completing the force correction. Simultaneously, the adjusted parameters are recorded and updated in the force control strategy library to guide subsequent grasping operations.

[0140] Through this dynamic force control correction mechanism, the adaptive fixture can automatically adjust the gripping force according to the actual situation of different blood gas consumables, ensuring both gripping firmness and avoiding damage to fragile consumables, thus significantly improving the adaptability and reliability of the automated blood gas analysis pretreatment system.

[0141] In one optional implementation, the spatial coordinates and placement posture constraint information of the placement area are extracted from the visual image, and the corresponding placement position is matched according to the consumable type identifier to generate a placement trajectory to the placement position, including:

[0142] The visual image is segmented into regions to identify the coordinates of the center point of the placement area, and the three-dimensional coordinates of the center point are used as the spatial position coordinates.

[0143] The placement direction and spacing of other blood gas consumables already placed in the placement area are identified from the visual image. The posture angle constraint range of the placement area is determined based on the placement direction, and the position offset constraint range of the placement area is determined based on the spacing, thereby obtaining the placement posture constraint information.

[0144] Based on the consumable type identifier, the priority placement order and preset placement slots of different types of blood gas consumables in the placement area are set to obtain placement rules; based on the placement rules, the spatial position coordinates and the placement posture constraint information, the corresponding placement position is determined;

[0145] The robot calculates the spatial displacement vector and attitude rotation vector between its current pose and the placement position. Based on the spatial displacement vector and attitude rotation vector, it plans a motion path to the placement position and sets corresponding timestamps and velocity parameters to generate the placement trajectory containing time and velocity information.

[0146] Deep learning-based image segmentation algorithms, such as U-Net or Mask R-CNN, are used to process the acquired visual images. Semantic segmentation distinguishes the placement region from the background. The centroid of the segmented placement region is calculated to obtain its two-dimensional planar coordinates (x, y). Combined with depth information from a depth camera, the depth value z of that point is obtained, thus determining the three-dimensional spatial coordinates (x, y, z) of the center point of the placement region. These coordinate values ​​directly serve as the spatial coordinates for subsequent placement operations, providing the robot with precise target location information.

[0147] The placement orientation and spacing of other blood gas analyzers already placed within the designated area are identified from visual images. Edge detection is performed on the existing blood gas analyzers within the area to extract their contours. Hough transform or principal component analysis is used to identify the dominant direction vectors, and the mean and standard deviation of these vectors are calculated to determine the dominant angle θ of the placement orientation and its fluctuation range Δθ. This angle range [θ-Δθ, θ+Δθ] constitutes the placement posture angle constraint range. Simultaneously, the Euclidean distance between adjacent analyzers is calculated, and the statistical distributions of the horizontal spacing dx and vertical spacing dy are obtained. Based on the mean and standard deviation of these spacing data, the constraint ranges for positional offset are determined: the lateral offset range [dx-σx, dx+σx] and the vertical offset range [dy-σy, dy+σy]. These angular and positional constraints together constitute the placement posture constraint information, ensuring that newly placed analyzers maintain a consistent arrangement with existing analyzers.

[0148] Based on the consumable type identifier, a priority placement order and preset placement slots are set for different types of blood gas consumables, establishing a mapping relationship between consumable types and placement rules. This includes: frequently used consumables are prioritized for easily accessible areas, emergency consumables are centrally placed in designated areas, and special consumables are stored in zones based on temperature sensitivity. For each type of consumable, a predefined placement slot matrix is ​​defined, with each slot containing a row and column index and a status marker (occupied / free). By querying this mapping relationship, the corresponding placement priority strategy and a list of available slots are obtained based on the consumable type identifier to be placed. Combining the previously obtained spatial coordinates and placement posture constraints, the optimal placement location that meets the constraints is selected from the available slots. The selection process considers the distance between the slot and the center point, the angle deviation, and the distribution density of surrounding consumables, ultimately determining the specific placement location coordinates and posture angle.

[0149] Obtain the current position coordinates (x1, y1, z1) and attitude quaternion q1 of the robot's end effector, and simultaneously determine the target placement position coordinates (x2, y2, z2) and attitude quaternion q2. Calculate the spatial displacement vector ΔP = (x2 - x1, y2 - y1, z2 - z1), representing the displacement from the current position to the target position. Calculate the attitude rotation vector ΔR, converting the current attitude quaternion q1 and the target attitude quaternion q2 into rotation matrices R1 and R2, and then calculate the relative rotation matrix Rrel = R2 · R1. -1Next, the relative rotation matrix Rrel is converted into an axis-angle representation, yielding the rotation axis n and rotation angle θ. The final attitude rotation vector ΔR is represented as the axis-angle product n·θ, which is the product of the rotation axis direction and the rotation angle magnitude, representing the rotation from the current attitude to the target attitude. Based on these vectors, a fifth-order polynomial interpolation algorithm is used to plan a smooth motion trajectory, avoiding sudden acceleration or deceleration during the robotic arm's movement. The entire motion process is discretized into several time points, each associated with corresponding position and attitude parameters, forming a timestamped pose sequence. Simultaneously, considering the fragility and stability requirements of the blood gas consumables, a suitable velocity curve is set to ensure smoothness and safety during movement. Finally, a complete placement trajectory containing time, position, attitude, and velocity information is generated for execution by the robot control system.

[0150] In practical applications, once a blood gas injector is identified as a standard type, a suitable empty space is found in the top left corner of the placement area using the method described above. The precise placement coordinates (125.3mm, 78.6mm, 15.2mm) and placement angle of 72° are calculated. A smooth trajectory is planned, allowing the robotic arm to start moving at an initial speed of 0.1m / s, with a maximum speed not exceeding 0.3m / s. Upon approaching the target position, the speed is reduced to 0.05m / s to complete the precise placement. The entire process takes approximately 2.5 seconds, ensuring the safe placement and neat arrangement of the blood gas injectors.

[0151] This method enables precise placement of blood gas consumables, ensuring the standardization and consistency of consumable arrangement within the placement area, and improving the automation level and efficiency of pre-analysis sample management in medical settings.

[0152] This invention relates to a visual positioning and placement device for mobile robotic blood gas analyzers, the device comprising:

[0153] The first unit is used to acquire visual images of blood gas consumables, storage areas, and placement areas through the robot's vision sensors;

[0154] The second unit is used to extract the geometric contour features of the blood gas consumable from the visual image, and construct the consumable pose model by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates and pose angle of the blood gas consumable.

[0155] The third unit is used to calculate the clamping distance and contact surface shape of the robot's adaptive gripper based on the size parameters and the consumable type identifier, and generate gripper configuration parameters.

[0156] The fourth unit is used to spatially associate the position of the blood gas consumable with the robot's current pose based on the three-dimensional spatial coordinates, and generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable.

[0157] The fifth unit is used to drive the robot to move to the storage area according to the collision-free motion trajectory, determine the approach direction based on the gripper configuration parameters and the posture angle, and generate a force control strategy based on the material characteristics of the blood gas consumable; during the grasping process, contact force feedback information is acquired in real time, and the force control strategy is dynamically corrected using the contact force feedback information;

[0158] The sixth unit is used to extract the spatial coordinates and placement posture constraint information of the placement area from the visual image, match the corresponding placement position according to the consumable type identifier, generate a placement trajectory to the placement position, and perform the placement action according to the placement trajectory.

[0159] A third aspect of the present invention provides an electronic device, comprising:

[0160] processor;

[0161] Memory used to store processor-executable instructions;

[0162] The processor is configured to invoke instructions stored in the memory to execute the aforementioned method.

[0163] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0164] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0165] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A machine vision-based intelligent positioning and placement method for blood gas analyzers, characterized in that, include: The robot's vision sensors acquire visual images of blood gas consumables, storage areas, and placement areas. The geometric contour features of the blood gas consumable are extracted from the visual image, and a consumable pose model is constructed by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates, and pose angle of the blood gas consumable, including: The pixel-level edge contour of the blood gas consumable in the visual image is extracted by an edge detection operator, and the pixel-level edge contour is fitted to obtain the two-dimensional geometric contour features of the blood gas consumable. Based on the two-dimensional geometric contour features, feature matching is performed to determine the consumable type identifier and size parameters of the blood gas consumables; The spatial coordinate system transformation relationship from the image coordinate system to the world coordinate system is established by using the intrinsic and extrinsic parameter matrices of the vision sensor. The pixel coordinates of feature points in the two-dimensional geometric contour features are extracted. Based on the spatial coordinate system transformation relationship, the pixel coordinates are converted into normalized coordinates in the camera coordinate system using the intrinsic parameter matrix. Then, the normalized coordinates are converted into the three-dimensional spatial coordinates in the world coordinate system using the extrinsic parameter matrix. Calculate the direction vector of the major axis direction in the two-dimensional geometric contour feature in the image coordinate system, and transform it to the world coordinate system through the spatial coordinate system transformation relationship. Based on the direction vector in the world coordinate system, calculate the rotation angle of the blood gas consumable about each coordinate axis of the world coordinate system to obtain the attitude angle of the blood gas consumable relative to the world coordinate system. Based on the size parameters and the consumable type identifier, the clamping distance and contact surface shape of the robot's adaptive gripper are calculated, and the gripper configuration parameters are generated. Based on the three-dimensional spatial coordinates, the position of the blood gas consumable is spatially correlated with the robot's current pose to generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable. The robot is driven to move to the storage area according to the collision-free motion trajectory, the approach direction is determined according to the fixture configuration parameters and the attitude angle, and a force control strategy is generated based on the material properties of the blood gas consumables. During the grasping process, contact force feedback information is acquired in real time, and the force control strategy is dynamically corrected using the contact force feedback information, including: During the gripping action of the adaptive clamp, the normal contact force value and tangential contact force value on the contact surface between the adaptive clamp and the blood gas consumable are collected in real time as the contact force feedback information. Extract the target contact force value corresponding to the current moment from the force control strategy, calculate the force deviation value between the normal contact force value and the target contact force value, and determine that a force correction operation needs to be performed when the force deviation value exceeds a preset force deviation threshold or the tangential contact force value exceeds a preset tangential force threshold. When performing force correction, the absolute value of the force deviation is taken as the force correction amplitude. The positive and negative directions of the force deviation are combined with the force correction amplitude to calculate the force correction amount and the corresponding direction of action. Based on the force correction amount, the clamping force output of the adaptive gripper is adjusted until the actual contact force value applied by the adaptive gripper converges to the target contact force value, thus obtaining the adjusted robot force control parameters. The adjusted robot force control parameters are then updated in the force control strategy. The spatial coordinates and placement posture constraint information of the placement area are extracted from the visual image, and the corresponding placement position is matched according to the consumable type identifier. A placement trajectory to the placement position is generated, and the placement action is performed according to the placement trajectory.

2. The method according to claim 1, characterized in that, Based on the size parameters and the consumable type identifier, the clamping distance and contact surface shape of the robot's adaptive gripper are calculated, and the gripper configuration parameters are generated, including: A three-dimensional geometric mesh for the blood gas consumable is constructed based on the size parameters, and the yield strength of the blood gas consumable is obtained according to the consumable type identifier. Multiple sets of candidate clamping configurations are set, and clamping force boundary conditions corresponding to each set of candidate clamping configurations are applied to the three-dimensional geometric mesh; Finite element analysis was performed on the three-dimensional geometric mesh after applying the clamping force boundary condition to obtain the stress distribution field and deformation distribution field of the blood gas consumable under each set of candidate clamping configurations. Extract the maximum stress value from the stress distribution field, extract the maximum deformation value from the deformation distribution field, and select the target clamping configuration from multiple candidate clamping configurations that makes the maximum stress value lower than the yield strength and the maximum deformation value the smallest. Extract the target clamping distance value and the target contact surface curvature value from the target clamping configuration, calculate the displacement control amount of the drive mechanism of the adaptive fixture based on the target clamping distance value, calculate the curvature control amount of the shape adjustment mechanism of the adaptive fixture based on the target contact surface curvature value, and encapsulate the displacement control amount and the curvature control amount into the fixture configuration parameters.

3. The method according to claim 1, characterized in that, Based on the three-dimensional spatial coordinates, the position of the blood gas consumable is spatially correlated with the robot's current pose to generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable, including: The robot's current position coordinates are used as the starting coordinates of the motion trajectory, and the three-dimensional spatial coordinates are used as the ending coordinates of the motion trajectory. In the world coordinate system, a spatial vector is constructed from the starting point coordinates to the ending point coordinates. The length of the spatial vector is calculated as the length of the motion path. Intermediate path nodes are set between the starting point coordinates and the ending point coordinates along the direction of the spatial vector. The number of intermediate path nodes is determined according to the length of the motion path. In the world coordinate system, a three-dimensional envelope of the space occupied by the obstacle is established. Each intermediate path node is determined in turn to see if it is located inside or on the boundary of the three-dimensional envelope. If so, the current intermediate path node is marked as a collision node. Calculate the shortest distance vector between the collision node and the surface of the three-dimensional envelope, and offset the position of the collision node in the opposite direction of the shortest distance vector to obtain the intermediate path node sequence after obstacle avoidance; By connecting the starting point coordinates, the sequence of intermediate path nodes after obstacle avoidance, and the ending point coordinates in chronological order, the collision-free motion trajectory is generated.

4. The method according to claim 2, characterized in that, Driving the robot to move to the storage area according to the collision-free motion trajectory, determining the approach direction based on the gripper configuration parameters and the attitude angle, and generating a force control strategy based on the material properties of the blood gas consumables includes: The collision-free motion trajectory is converted into an angle command sequence for each joint of the robot, and the angle commands of each joint in the angle command sequence are executed sequentially in chronological order to drive the robot to move to the storage area; Based on the fixture configuration parameters, the contact normal vector when the adaptive fixture contacts the blood gas consumable is calculated according to the target clamping distance value and the target contact surface curvature value, and the direction of the contact normal vector in the world coordinate system is determined as the approach direction. Obtain the corresponding material hardness value and compressive strength value according to the consumable type identifier, calculate the maximum contact force threshold that can be applied during the gripping process based on the material hardness value, and calculate the maximum deformation threshold that can be generated during the gripping process based on the compressive strength value; The maximum contact force threshold is calculated based on the material hardness value, and a contact force control curve is generated. The maximum deformation threshold is calculated based on the compressive strength value, and a deformation monitoring threshold is set. The force control strategy is set according to the approach direction, the contact force control curve, and the deformation monitoring threshold.

5. The method according to claim 1, characterized in that, The spatial coordinates and placement posture constraint information of the placement area are extracted from the visual image, and the corresponding placement position is matched according to the consumable type identifier to generate a placement trajectory to the placement position, including: The visual image is segmented into regions to identify the coordinates of the center point of the placement area, and the three-dimensional coordinates of the center point are used as the spatial position coordinates. The placement direction and spacing of other blood gas consumables already placed in the placement area are identified from the visual image. The posture angle constraint range of the placement area is determined based on the placement direction, and the position offset constraint range of the placement area is determined based on the spacing, thereby obtaining the placement posture constraint information. Based on the consumable type identifier, the priority placement order and preset placement slots of different types of blood gas consumables in the placement area are set to obtain the placement rules; The corresponding placement position is determined based on the placement rules, the spatial coordinates, and the placement posture constraint information; The robot calculates the spatial displacement vector and attitude rotation vector between its current pose and the placement position. Based on the spatial displacement vector and attitude rotation vector, it plans a motion path to the placement position and sets corresponding timestamps and velocity parameters to generate the placement trajectory containing time and velocity information.

6. A machine vision-based intelligent positioning and placement system for blood gas analyzers, used to implement the method as described in any one of claims 1-5, characterized in that, include: The first unit is used to acquire visual images of blood gas consumables, storage areas, and placement areas through the robot's vision sensors; The second unit is used to extract the geometric contour features of the blood gas consumable from the visual image, and construct the consumable pose model by combining spatial coordinate system transformation to obtain the consumable type identifier, size parameters, three-dimensional spatial coordinates and pose angle of the blood gas consumable. The third unit is used to calculate the clamping distance and contact surface shape of the robot's adaptive gripper based on the size parameters and the consumable type identifier, and generate gripper configuration parameters. The fourth unit is used to spatially associate the position of the blood gas consumable with the robot's current pose based on the three-dimensional spatial coordinates, and generate a collision-free motion trajectory from the current pose to the position of the blood gas consumable. The fifth unit is used to drive the robot to move to the storage area according to the collision-free motion trajectory, determine the approach direction based on the fixture configuration parameters and the attitude angle, and generate a force control strategy based on the material properties of the blood gas consumables. During the grasping process, contact force feedback information is acquired in real time, and the force control strategy is dynamically corrected using the contact force feedback information; The sixth unit is used to extract the spatial coordinates and placement posture constraint information of the placement area from the visual image, match the corresponding placement position according to the consumable type identifier, generate a placement trajectory to the placement position, and perform the placement action according to the placement trajectory.

7. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 5.

Images