Method and system for regulation optimization of all-dielectric intelligent sampling robot

Abstract

The application provides a control optimization method and system for a full-medium intelligent sampling robot, and relates to the technical field of robot control. The method comprises the following steps: analyzing a sampling medium type, controlling a multi-medium adaptive end effector to rotate and switch to a corresponding sampling head station and lock, constructing a virtual gauge plate space model, generating a standardized sampling path, performing a sampling action, simultaneously collecting real-time pose data and contact force feedback data of a mechanical arm, detecting a disinfectant concentration in a self-disinfection process, adaptively optimizing and adjusting sampling standard parameters based on the pose data, the contact force feedback data and disinfection effect verification results, and obtaining adjusted sampling parameters. The technical problems that the prior art cannot accurately adapt to various sampling media, and the path and disinfection control are extensive, resulting in insufficient sampling accuracy and reliability are solved. The technical effects of accurately adapting to different sampling media, efficiently sampling and reliably self-disinfecting are achieved.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and more specifically to a method and system for regulating and optimizing an all-media intelligent sampling robot. Background Technology

[0002] With the development of public health monitoring, environmental testing, and biosafety, sampling robots are increasingly being used for sample collection from various media, including air, water, and object surfaces. Compared to manual sampling, robotic sampling offers advantages such as high automation, good repeatability, and the ability to operate in complex or hazardous environments.

[0003] However, existing sampling robots are mostly designed and planned for single media, lacking the ability to adaptively switch between different sampling media, making it difficult to achieve unified application across multiple scenarios. Furthermore, in terms of sampling path generation, they often rely on preset fixed trajectories or simple planar assumptions, failing to dynamically model and calibrate based on the spatial geometric characteristics of the actual sampling area, which can easily lead to sampling point deviations or uneven coverage. In addition, existing disinfection processes are mostly timed or quantitatively controlled, lacking real-time detection and closed-loop verification of disinfectant concentration and disinfection effectiveness, posing a risk of cross-contamination and thus affecting sampling accuracy and reliability.

[0004] Existing technologies have technical problems such as difficulty in accurately adapting to diverse sampling media and crude path and disinfection control, resulting in insufficient sampling accuracy and reliability. Summary of the Invention

[0005] The purpose of this application is to provide a control and optimization method and system for a full-media intelligent sampling robot, which solves the technical problems of existing technologies that are difficult to accurately adapt to diverse sampling media and have crude path and disinfection control, resulting in insufficient sampling accuracy and reliability.

[0006] In view of the above problems, this application provides a method and system for regulating and optimizing an all-media intelligent sampling robot.

[0007] The first aspect of this application provides a control and optimization method for an all-media intelligent sampling robot. The method includes: receiving a sampling task instruction; parsing the sampling medium type; controlling a multi-media adaptive end effector to rotate and switch to the corresponding sampling head position and lock it according to the target medium type; constructing a virtual specification board space model through infrared ranging and fluorescence projection; performing three-dimensional modeling and boundary calibration of the target sampling area; and generating a standardized sampling path according to the sampling standard parameters in the sampling task instruction; controlling a multi-joint robotic arm to perform sampling actions according to the standardized sampling path, while simultaneously collecting the robotic arm's pose data and contact force feedback data in real time, and executing an in-situ self-disinfection process before and after sampling; detecting the disinfectant concentration in the self-disinfection process and generating a disinfection effect verification result; and adaptively optimizing and adjusting the sampling standard parameters based on the pose data, contact force feedback data, and disinfection effect verification result to obtain adjusted sampling parameters.

[0008] Optionally, the sampling medium type includes air medium, object surface medium, and water medium; after obtaining the target medium type, the multi-medium adaptive end effector is controlled to rotate and switch to the corresponding plate holder, swab holder, or water sample collection needle module.

[0009] Optionally, the infrared ranging fluorescence projection module is controlled to project an array of structured light spots onto the target sampling area, and a vision sensor set at the end of the robotic arm is used to acquire the distorted image of the structured light spots in physical space; a three-dimensional reconstruction algorithm is executed based on the distorted image to obtain spatial point cloud data of the target sampling area, a virtual specification board spatial model is constructed, and the boundary calibration of the sampling area is completed; according to the sampling standard parameters in the sampling task instruction, the logical point sequence of the virtual specification board is mapped to the spatial model to generate a standardized sampling path containing spatial position and attitude information.

[0010] Optionally, the virtual specification board is dynamically parameterized and adjusted according to the sampling standard parameters in the sampling task instruction to obtain the adjusted virtual specification board; the end effector of the robotic arm carries the calibration board to traverse the corner positions of the adjusted virtual specification board, and the actual spatial coordinates of the calibration board are synchronously collected by the vision sensor to obtain the logical point sequence; the rigid transformation matrix from the projected coordinate system to the robotic arm base coordinate system is solved online by the least squares method; the logical point sequence is substituted into the rigid transformation matrix to generate the target pose sequence, and the target pose sequence is optimized and sorted based on the shortest path principle to obtain the standardized sampling path.

[0011] Optionally, the disinfectant concentration is detected at a preset detection frequency to obtain a disinfectant concentration sequence; it is determined whether the last disinfectant concentration in the disinfectant concentration sequence is less than a preset disinfectant concentration threshold. If so, a concentration trend analysis command is triggered; based on the concentration trend analysis command, a trend analyzer is called to perform trend analysis on the disinfectant concentration sequence, and the trend analysis result and the result of incomplete disinfection are used as the disinfection effect verification result; otherwise, the result of completed disinfection is used as the disinfection effect verification result.

[0012] Optionally, trajectory deviation analysis is performed on the pose data to obtain position error and attitude error; stability analysis is performed on the contact force feedback data to obtain contact pressure fluctuation parameters; based on the position error, attitude error, contact pressure fluctuation parameters and disinfection effect verification results, the sampling standard parameters are adaptively adjusted according to preset optimization rules to obtain adjusted sampling parameters.

[0013] Optionally, the sampling standard parameters include sampling point spacing, pressing depth, dwell time, and path smoothing coefficient.

[0014] A second aspect of this application provides a control and optimization system for an all-media intelligent sampling robot. This system includes: a media type analysis component, used to receive sampling task instructions, analyze the sampling media type, and control a multi-media adaptive end effector to rotate and switch to the corresponding sampling head position and lock it according to the target media type; a sampling path generation component, used to construct a virtual specification board space model through infrared ranging and fluorescence projection, perform three-dimensional modeling and boundary calibration of the target sampling area, and generate a standardized sampling path according to the sampling standard parameters in the sampling task instructions; a data acquisition component, used to control a multi-joint robotic arm to perform sampling actions according to the standardized sampling path, while simultaneously acquiring the robotic arm's pose data and contact force feedback data in real time, and executing an in-situ self-disinfection process before and after sampling; a concentration detection component, used to detect the disinfectant concentration in the self-disinfection process and generate disinfection effect verification results; and a sampling parameter adjustment component, used to adaptively optimize and adjust the sampling standard parameters based on the pose data, contact force feedback data, and disinfection effect verification results to obtain adjusted sampling parameters.

[0015] One or more technical solutions provided in this application have at least the following technical effects or advantages: The method provided in this application receives a sampling task instruction, parses the sampling medium type, and controls a multi-media adaptive end effector to rotate and switch to the corresponding sampling head position and lock it according to the target medium type. A virtual specification plate space model is constructed using infrared ranging and fluorescence projection to perform 3D modeling and boundary calibration of the target sampling area, and a standardized sampling path is generated according to the sampling standard parameters in the sampling task instruction. A multi-joint robotic arm is controlled to perform sampling actions according to the standardized sampling path, while simultaneously collecting the robotic arm's pose data and contact force feedback data in real time, and executing an in-situ self-disinfection process before and after sampling. The disinfectant concentration in the self-disinfection process is detected, and a disinfection effect verification result is generated. Based on the pose data, contact force feedback data, and disinfection effect verification result, the sampling standard parameters are adaptively optimized and adjusted to obtain adjusted sampling parameters. This achieves the technical effect of accurately adapting to different sampling media, realizing efficient sampling and reliable self-disinfection, thereby improving the accuracy and reliability of sampling.

[0016] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1 A flowchart illustrating the control and optimization method for the all-media intelligent sampling robot provided in this application.

[0019] Figure 2 This is a schematic diagram of the control and optimization system for the all-media intelligent sampling robot provided in this application.

[0020] Explanation of reference numerals in the attached diagram: 11. Medium type analysis component; 12. Sampling path generation component; 13. Data acquisition component; 14. Concentration detection component; 15. Sampling parameter adjustment component. Detailed Implementation

[0021] This application provides a control and optimization method and system for an intelligent sampling robot for all media, addressing the technical problems of existing technologies that struggle to accurately adapt to diverse sampling media and suffer from coarse path and disinfection control, resulting in insufficient sampling accuracy and reliability. It achieves precise adaptation to different sampling media, enabling efficient sampling and reliable self-disinfection, thereby improving the accuracy and reliability of sampling.

[0022] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0023] Example 1, as Figure 1 As shown, this application provides a control and optimization method for an all-media intelligent sampling robot, the control and optimization method for the all-media intelligent sampling robot includes: Receive sampling task instructions, parse the sampling medium type, and control the multi-media adaptive end effector to rotate and switch to the corresponding sampling head position and lock it according to the target medium type.

[0024] Furthermore, the sampling medium type includes air medium, object surface medium, and water medium; after obtaining the target medium type, the multi-medium adaptive end effector is controlled to rotate and switch to the corresponding plate holder, swab holder, or water sample collection needle module.

[0025] Specifically, after receiving a sampling task instruction from the operator terminal or central control system, the sampling robot first parses the data structure of the instruction. This instruction typically uses a structured data format, such as a JSON data frame or a predefined field protocol message, which includes a task number, sampling area number, sampling standard parameters, and a media type field. The robot then unpacks the data message and performs keyword matching to extract the value or enumeration code of the media type identifier field. For example, 01 represents air, 02 represents object surface, and 03 represents water. This enumeration code is converted into the corresponding sampling media type using a pre-defined media type mapping table for the end effector to call. Sampling media types include air, object surface, and water. Air is suitable for gas sampling, such as air quality monitoring; object surface is suitable for sampling solid surfaces, such as collecting microorganisms or pollutants; and water is used for water sample collection, suitable for water body detection or pollution source tracing.

[0026] Based on the identified target medium type, the rotation of the multi-media adaptive end effector is controlled to switch it to the corresponding sampling head. The multi-media adaptive end effector is a flexible, modular robotic arm integrating multiple independently rotatable modules. Each module corresponds to a sampling head: a petri dish holder, a swab holder, or a water sampling needle. The sampling tools can be changed as needed, allowing the sampling robot to adapt to different sampling requirements. The petri dish holder is primarily used for collecting microbial samples from air, precisely holding the petri dish to ensure its stability and sealing during sampling. The swab holder is for surface media, firmly fixing the swab so it can effectively contact the surface and collect samples. The water sampling needle is used for water sampling; its slender needle-like structure can penetrate deep into the water to accurately collect samples.

[0027] After receiving the control command, the end effector drives the rotating mechanism to rotate at a predetermined angle and speed through an internal drive device, such as a high-precision stepper motor. This rotates the corresponding sampling head to the working position, and a reliable locking mechanism, such as an electromagnetic lock or mechanical latch, securely locks the sampling head to prevent it from shaking or shifting during sampling, thus ensuring the accuracy and stability of the sampling.

[0028] By dynamically switching to the appropriate sampling tool based on the type of sampling medium, it ensures that different types of media can be sampled in the most suitable way, enabling the robot to quickly adapt to various sampling tasks. This not only improves sampling efficiency but also prevents sampling errors or contamination caused by tool incompatibility, thereby ensuring the accuracy and reliability of sampling results.

[0029] A virtual specification board spatial model is constructed by infrared ranging and fluorescence projection. The target sampling area is then three-dimensionally modeled and its boundaries are calibrated. A standardized sampling path is generated based on the sampling standard parameters in the sampling task instruction.

[0030] Furthermore, a virtual specification board spatial model is constructed using infrared ranging and fluorescence projection. The target sampling area is then 3D modeled and its boundaries calibrated. A standardized sampling path is generated based on the sampling standard parameters in the sampling task instruction. This includes: controlling the infrared ranging and fluorescence projection module to project an array of structured light spots onto the target sampling area, and acquiring distorted images of the structured light spots in physical space using a vision sensor located at the end of the robotic arm; executing a 3D reconstruction algorithm based on the distorted images to obtain spatial point cloud data of the target sampling area, constructing a virtual specification board spatial model, and completing boundary calibration of the sampling area; and mapping the logical point sequence of the virtual specification board to the spatial model according to the sampling standard parameters in the sampling task instruction, generating a standardized sampling path containing spatial position and attitude information.

[0031] Furthermore, the sampling standard parameters include sampling point spacing, pressing depth, dwell time, and path smoothing coefficient.

[0032] Specifically, after the sampling robot completes the identification of the sampling medium type and the switching of the sampling head, it controls the infrared ranging and fluorescence projection module to project an array of structured light spots onto the target sampling area. The array of structured light spots refers to a high-density set of light points arranged according to preset geometric rules, such as a rectangular grid or concentric circle array, used to form a calculable spatial reference mark on the measured physical surface. Infrared ranging is used to obtain distance information of the target area, and fluorescence projection is used to enhance surface reflection contrast, ensuring the uniformity and stability of the light spots. The projected light spots can cover the entire target sampling area. Simultaneously, a vision sensor located at the end of the robotic arm, such as an industrial camera or a depth vision module, synchronously acquires distorted images of the structured light spots in actual space due to the different shapes of the target object's surface. These distorted images refer to the geometric deformation images of the structured light spots caused by surface undulations or tilts, reflecting the concavity, tilt, and other features of the target object's surface, and are used to obtain the object's three-dimensional information.

[0033] After acquiring the distorted image formed by the structured light array on the target surface, the distorted image is first converted into a grayscale image. A weighted averaging method, such as Gray = 0.299×R + 0.587×G + 0.114×B, is used to reduce data dimensionality and enhance the stability of brightness information. The grayscale image is then denoised, preferably using a Gaussian filtering algorithm. By setting the convolution kernel size (e.g., 5×5) and standard deviation σ, the grayscale image is smoothed to suppress ambient light interference and sensor random noise while maintaining the continuity of the light spot edges. After filtering, the preprocessed distorted image is binarized using adaptive thresholding or the Otsu algorithm to extract the bright structured light regions. Connected component analysis is then used to identify each independent light spot region. For each light spot region, the centroid method is used to calculate its pixel-level center coordinates (u, v), where u represents the horizontal pixel coordinate and v represents the vertical pixel coordinate. This involves weighted summation of the grayscale values ​​of all pixels within the light spot region to obtain the initial center position. Combining the known calibration parameters between the infrared ranging fluorescence projection module and the vision sensor, including the intrinsic parameter matrix, extrinsic parameter matrix, and baseline distance, a 3D reconstruction algorithm based on the triangulation principle is invoked. Based on the offset between the theoretical projection position of the light spot and the actual pixel displacement, the depth value Z of the corresponding spatial point is calculated. Then, the spatial 3D coordinates (X,Y,Z) are calculated using the back projection formulas X=Z·(u-cx) / fx and Y=Z·(v-cy) / fy to generate spatial point cloud data of the target sampling area. Here, fx and fy refer to the focal length parameters in the camera's intrinsic parameter matrix, representing the equivalent focal length in the x and y directions, respectively. cx and cy refer to the coordinates of the camera's principal point, i.e., the projection center position of the optical axis on the image plane. After acquiring the spatial point cloud data of the target sampling area, the least squares method is used to fit a plane or quadratic surface to obtain the surface equation parameters. The flatness is calculated by statistical analysis of the residual of the fitted plane. The flatness is the difference between the maximum and minimum residuals, which characterizes the surface undulation error. The tilt angle θ = arccos(n·Z0 / |n|) is calculated by the angle between the normal vector of the fitted plane and the Z-axis of the robot arm base coordinate system, where n is the normal vector of the fitted plane, Z0 is the unit vector of the Z-axis in the robot arm base coordinate system, θ is the tilt angle of the sampling surface relative to the robot arm base coordinate system, and |n| is the magnitude of the normal vector n. If it is a curved surface, the principal curvature and radius of curvature R = 1 / k are calculated based on the quadratic surface parameters. The larger R is, the closer it is to a plane. k is the principal curvature of the surface, which represents the degree of curvature per unit length.

[0034] After obtaining the geometric feature parameters of the sampling surface, including flatness, tilt angle, and radius of curvature, the surface analytical expression is obtained based on the least squares fitting result of the point cloud, such as the plane equation Ax+By+Cz+D=0 or the quadratic surface equation. The surface analytical expression is used as the spatial reference surface of the virtual specification board. The flatness parameter is used to evaluate the maximum positive and negative residual range from the point cloud to the fitting surface. The outer contour point corresponding to the extreme value of the residual is used as the height compensation boundary of the sampling area. The tilt angle is determined by the angle between the fitting normal vector and the Z-axis of the robot arm base coordinate system, thereby establishing the attitude rotation matrix of the virtual specification board relative to the robot arm base coordinate system to achieve model attitude alignment. If curvature is detected, a surface parameterized coordinate system is constructed according to the principal curvature direction. Under the condition of continuous change of the surface normal, an adaptively fitted surface-type virtual specification board is generated. The spatial point cloud data of the target sampling area is projected in two dimensions onto a fitting reference plane or surface parameter domain. The outer contour boundary of the point cloud is calculated using a convex hull algorithm or an α-shape boundary extraction algorithm to determine the effective geometric range of the sampling area. This boundary is then used as a clipping constraint for the virtual specification board. Finally, a virtual specification board spatial model that matches the real sampling surface in position, orientation, and shape is generated, achieving boundary calibration of the sampling area. The virtual specification board spatial model is a digital representation of the target sampling area that can accurately reflect its spatial structure and characteristics.

[0035] Based on the sampling standard parameters in the sampling task instruction, including sampling point spacing, pressing depth, dwell time and path smoothing coefficient, the virtual specification board is parametrically scaled and its density adjusted. The logical point sequence is mapped to the actual spatial model through the coordinate transformation matrix, so that the logical sampling points are aligned with the geometric features of the real sampling surface, thereby generating a standardized sampling path containing spatial position and attitude information, namely three-dimensional coordinates and end-effector normal attitude angle.

[0036] By using infrared ranging and fluorescence projection technology to construct a virtual specification board spatial model, the target sampling area is accurately modeled in three dimensions and its boundaries are calibrated. A standardized sampling path is generated to ensure that the robotic arm of the sampling robot samples according to an accurate and reasonable path, thereby improving the efficiency and accuracy of sampling, reducing sampling errors, and ensuring the reliability and consistency of the sampling results.

[0037] Furthermore, based on the sampling standard parameters in the sampling task instruction, the logical point sequence of the virtual specification board is mapped to the spatial model to generate a standardized sampling path containing spatial position and attitude information. This includes: dynamically parameterizing the virtual specification board according to the sampling standard parameters in the sampling task instruction to obtain an adjusted virtual specification board; controlling the end effector of the robotic arm to carry the calibration board and traverse the corner positions of the adjusted virtual specification board, and synchronously collecting the actual spatial coordinates of the calibration board by a vision sensor to obtain a logical point sequence; solving the rigid transformation matrix from the projected coordinate system to the robotic arm's base coordinate system online using the least squares method; substituting the logical point sequence into the rigid transformation matrix to generate a target pose sequence, and optimizing and sorting the target pose sequence based on the shortest path principle to obtain a standardized sampling path.

[0038] Specifically, the virtual specification board is dynamically and parametrically adjusted according to the sampling standard parameters in the sampling task instruction. The sampling standard parameters are preset and are a series of indicators used to standardize the sampling process, including sampling point spacing, path smoothness coefficient, pressing depth, and dwell time. Among them, sampling point spacing refers to the minimum distance between adjacent sampling points in space, which is used to ensure uniform and non-repetitive sampling coverage. Path smoothness coefficient is used to control the continuity and smoothness of the robotic arm's trajectory on the sampling path, reducing sharp turns and vibrations. Pressing depth refers to the vertical pressing depth when the end sampling tool contacts the sampling surface, ensuring sufficient sample collection without damaging the surface. Dwell time refers to the length of time the sampling tool stays at each sampling point, which is used to ensure complete sample collection. Based on these sampling standard parameters, the logical points of the virtual specification board are first scaled and redistributed at equal intervals according to the sampling point spacing and the actual size of the sampling area. Then, translation and rotation adjustments are made according to the sampling surface tilt angle and boundary shape to make the virtual specification board fit the actual spatial shape of the target sampling area. At the same time, the connection sequence of the points is initially optimized by combining the path smoothing coefficient to generate a continuous and smooth sequence of logical sampling points. Pressing depth and dwell time are bound as additional attributes to each logical point to form the adjusted virtual specification board, thus providing an accurate and executable spatial reference for subsequent robotic arm path planning and standardized sampling. For example, setting the sampling point spacing to 20mm represents the minimum spacing between adjacent sampling points on the virtual specification board to ensure uniform coverage and no repetition. If the target sampling area is 100mm×80mm, 5 points can be arranged in the length direction and 4 points in the width direction, for a total of 20 sampling points. The path smoothness coefficient is set to 0.8, ranging from 0 to 1. A value closer to 1 indicates a smoother path, used to smoothly connect adjacent points when generating the sampling path, reducing sharp turns and acceleration changes in the robotic arm, and ensuring stable sampling movements. The pressing depth is set to 5mm, representing the depth to which the sampling tool, such as a swab or sampling needle, presses into the surface when it contacts the sampling surface, ensuring sufficient sample collection without damaging the surface. The dwell time is set to 2s, meaning the sampling tool stays at each sampling point for 2 seconds, ensuring sufficient contact and sample adsorption. That is, when the robotic arm samples on a 100mm × 80mm inclined surface, it generates 5 × 4 logical sampling points based on a 20mm point spacing. Combined with the path smoothness coefficient of 0.8, the path sequence is optimized, making the robotic arm move smoothly. At the end of each sampling point, the actuator presses 5mm and stays for 2 seconds before moving to the next point, achieving uniform, standardized, and high-precision sampling operations.

[0039] The robotic arm's end effector, carrying a calibration plate, traverses the corner positions of the adjusted virtual specification board. Simultaneously, the end effector's vision sensor acquires the actual 3D spatial coordinates of the calibration plate, obtaining a sequence of logical points corresponding to the logical points on the virtual specification board. The calibration plate serves as a standard reference for calibrating spatial coordinates and attitude. The logical point sequence is represented in 3D coordinates within the virtual specification board coordinate system. The correspondence between logical points and actual points is established using the actual spatial coordinates acquired by the robotic arm's end effector calibration plate. The least squares method is used to solve the rigid transformation matrix online. This rigid transformation matrix contains a rotation matrix R (3×3) and a translation vector T (3×1), which can transform any point P in the virtual specification board coordinate system. v =[x v ,y v ,z v ] T Mapped to the actual spatial point P in the robot arm's base coordinate system r =R×P v +T. Substituting the logical point sequence into the rigid transformation matrix one by one, the target spatial pose sequence is obtained. Each pose includes: position, attitude, pressing depth, and dwell time. Among them, position refers to the spatial coordinates (X,Y,Z) of the end effector in the robot arm base coordinate system. The attitude is the normal attitude angle of the end tool, which is calculated based on the normal vector of the virtual specification plate and the normal vector of the fitted surface. It is used to ensure that the sampling tool is perpendicular to or fits the target sampling surface. The pressing depth and dwell time are derived from the sampling standard parameters and are used to control the depth and duration of the sampling tool pressing into the surface, respectively.

[0040] The target pose sequence is optimized and sorted according to the shortest path principle to improve the efficiency of the robotic arm movement and ensure sampling accuracy. Specifically, an Euclidean distance matrix is ​​established between each pose to quantify the movement cost of the robotic arm's end effector in space, while the cost of sharp turns is calculated in conjunction with posture changes. Optimization algorithms, such as greedy algorithms or nearest neighbor algorithms, are invoked to iteratively select the next sampling point with the shortest distance to the current position and satisfying row / column scanning or a near-to-far order logic, starting from the starting point. These sampling points are then connected sequentially to form a preliminary path. To ensure path smoothness, the preliminary path is further interpolated and filtered to adjust the continuous angle changes between poses, preventing excessive acceleration or sharp turns during the robotic arm's end effector movement. After optimization, each pose not only retains its original spatial position, posture, pressing depth, and dwell time, but also minimizes the total movement distance and reduces the number of turns in the sequence, achieving a standardized sampling path and ensuring sampling uniformity, efficiency, and stability.

[0041] By dynamically parameterizing and adjusting the virtual specification board to adapt to the needs of different sampling tasks, collecting actual spatial coordinates and solving the rigid transformation matrix, a precise mapping between virtual and actual space is achieved. The target pose sequence is generated and optimized, providing a scientific and reasonable motion path for the robotic arm, realizing high-precision, uniform, and continuous standardized sampling actions, thereby ensuring the accuracy and efficiency of sampling.

[0042] The multi-joint robotic arm is controlled to perform sampling actions according to the standardized sampling path, while the positional data and contact force feedback data of the robotic arm are collected in real time, and an in-situ self-disinfection process is performed before and after sampling.

[0043] Specifically, after obtaining the standardized sampling path, the multi-joint robotic arm is driven to sequentially execute sampling actions according to the standardized sampling path. The multi-joint robotic arm refers to a robot arm with multiple rotational or translational degrees of freedom, which can precisely control the position and attitude of the end effector in three-dimensional space. During the sampling process, data from the encoders or absolute rotation sensors of each joint are continuously read through a real-time communication interface, such as EtherCAT or CAN bus, to obtain the angle and angular velocity of each joint. Then, using the forward kinematics model of the robotic arm, the joint angles are mapped to the three-dimensional position and attitude of the end effector in the robotic arm's base coordinate system, realizing real-time acquisition of pose data. Simultaneously, a high-precision six-degree-of-freedom force or torque sensor is installed between the end effector and the sampling tool. The force / torque signal is acquired in real time to the control system through a high-speed analog-to-digital converter, recording the normal force, tangential force, and torque changes of the end contact surface. This data is used to evaluate contact stability and control the pressing depth, realizing real-time acquisition of contact force feedback data. To prevent cross-contamination, an in-situ self-disinfection process is performed before and after each sampling. In-situ self-disinfection means that the robotic arm does not need to move to a separate disinfection area, but completes the disinfection operation at the current sampling position or near the end. It covers the surface of the sampling tool by spraying or atomizing disinfectant, and works with a disinfectant concentration detection sensor to verify in real time whether the disinfection effect has reached the preset threshold.

[0044] During the sampling process, the robot arm's pose data and contact force feedback data are collected in real time, which can promptly understand the robot arm's motion status and the contact situation during the sampling process, ensuring that the sampling action is accurate, uniform, and stable. Furthermore, the in-situ self-disinfection mechanism maintains the sterility of the sampling environment and tools by performing in-situ self-disinfection procedures before and after sampling, effectively preventing cross-contamination and thus ensuring the reliability and safety of the sampling results.

[0045] The concentration of disinfectant in the self-disinfection process is detected to generate disinfection effect verification results.

[0046] Furthermore, the disinfectant concentration in the self-disinfection process is detected to generate a disinfection effect verification result, including: detecting the disinfectant concentration at a preset detection frequency to obtain a disinfectant concentration sequence; determining whether the last disinfectant concentration in the disinfectant concentration sequence is less than a preset disinfectant concentration threshold; if so, triggering a concentration trend analysis command; based on the concentration trend analysis command, calling a trend analyzer to perform trend analysis on the disinfectant concentration sequence, and using the trend analysis result and incomplete disinfection as the disinfection effect verification result; otherwise, using completed disinfection as the disinfection effect verification result.

[0047] Specifically, the disinfectant concentration is detected in real time by a disinfectant concentration sensor located at the end of the device or near the spraying unit. This sensor can employ optical or electrochemical sensing technologies to quantify the concentration of disinfectant in the air or liquid. Optical sensing methods include light absorption or fluorescence, while electrochemical sensing methods include ion-selective electrodes. Sensor data is continuously read at a preset detection frequency, for example, once per second, generating a disinfectant concentration sequence and recording the concentration change trend during the disinfection process.

[0048] The concentration of the last disinfectant in the disinfectant concentration sequence is compared with a preset disinfectant concentration threshold, which is the minimum safe and effective concentration pre-set according to actual disinfection specifications, such as 5%. If the concentration of the last disinfectant in the sequence is less than the preset threshold, it indicates that the current disinfectant coverage concentration has not yet met the effective sterilization requirements, the disinfection process may not be complete, or the concentration is still rising. Therefore, a concentration trend analysis command is triggered, calling the built-in trend analyzer to evaluate the trend of the entire concentration sequence to determine the direction of concentration change, growth rate, and stability. The trend analyzer can use a moving average method to analyze the disinfectant concentration sequence data. First, a fixed-length moving window is set, for example, selecting disinfectant concentration data from three consecutive time points as a window. Then, the average value of the data within this window is calculated as the representative value of the disinfectant concentration during this period. As the data is continuously updated, the moving window moves sequentially forward in time, and the average value of the data within the new window is recalculated each time it moves. By comparing the average values ​​of adjacent windows, if the average value of the new window is greater than that of the previous window, it indicates an upward trend in disinfectant concentration. If the average value of the new window is less than that of the previous window, it indicates a downward trend in disinfectant concentration. If the difference between the average values ​​of adjacent windows fluctuates within a very small range, it can be considered that the disinfectant concentration is in a stable trend. Based on the trend analysis results, the disinfection status is marked as incomplete, prompting continued spraying or an extension of the disinfection time. If the disinfectant concentration at the last point is greater than or equal to the preset disinfectant concentration threshold, it means that the disinfectant concentration in the current disinfection area has reached or exceeded the minimum concentration standard required for safe and effective sterilization. The disinfectant has formed an effective coverage on the surface of the sampling tool or the target area that meets the specifications, and has the conditions to achieve the expected sterilization or inactivation effect. In this case, the current round of self-disinfection process is judged to have reached a qualified state, and trend extension analysis or re-spray control commands are no longer triggered. The completion of disinfection is taken as the disinfection effect verification result.

[0049] By monitoring and analyzing the concentration of disinfectants in real time, we can understand the changes in the concentration of disinfectants during the disinfection process, accurately determine whether the disinfection has achieved the expected effect, ensure that the end tools and the surface of the sampling area are fully disinfected before and after each sampling, reduce the risk of cross-contamination, and thus ensure the reliability and safety of the sampling data.

[0050] Based on the pose data, contact force feedback data, and disinfection effect verification results, the sampling standard parameters are adaptively optimized and adjusted to obtain the adjusted sampling parameters.

[0051] Furthermore, based on the pose data, contact force feedback data, and disinfection effect verification results, the sampling standard parameters are adaptively optimized and adjusted to obtain the adjusted sampling parameters. This includes: performing trajectory deviation analysis on the pose data to obtain position error and posture error; performing stability analysis on the contact force feedback data to obtain contact pressure fluctuation parameters; and adaptively adjusting the sampling standard parameters according to preset optimization rules based on the position error, posture error, contact pressure fluctuation parameters, and disinfection effect verification results to obtain the adjusted sampling parameters.

[0052] Specifically, trajectory deviation analysis is performed based on the acquired pose data. This involves comparing the actual pose of the end effector in three-dimensional space with the target pose of the standardized sampling path, calculating the position error and attitude error at each sampling point to assess the degree of trajectory deviation. The position error is the Euclidean distance between the actual and target positions, and the attitude error is the angle difference between the actual and target attitudes. Simultaneously, stability analysis is performed on the contact force feedback data acquired by the end effector. Statistics such as the standard deviation and variance of the contact force feedback data are calculated to obtain contact pressure fluctuation parameters, which are used to assess the stability of the pressing depth and applied force during the sampling process.

[0053] Then, combining the results of positional error, contact force fluctuation, and disinfection effect verification, the sampling standard parameters are adaptively adjusted according to preset optimization rules to obtain adjusted sampling parameters. Adjusted sampling parameters refer to the set of parameters resulting from the dynamic correction and updating of the original sampling standard parameters. For example, if the positional error exceeds the posture deviation threshold, the path smoothing coefficient is increased to reduce trajectory jitter; if the contact pressure fluctuation is too large, the pressing depth or the end-effector speed is adjusted to improve sampling uniformity; if disinfection is incomplete, the dwell time is extended or the disinfection spray frequency is adjusted. The positional deviation threshold is set according to the positioning accuracy and sampling accuracy requirements of the robotic arm. For example, the positional error threshold is set to 0.5-1mm, and the posture error threshold is set to 1-2° to ensure that the sampling point deviation is within an acceptable range. The contact pressure fluctuation threshold is set according to the allowable force fluctuation range of the sampling surface and the force that the sampling tool can withstand. For example, the average normal force of swab sampling is 1.2N, and a fluctuation of ±0.3N is allowed. If the standard deviation exceeds 0.3N, it indicates that the force application is unstable, and the pressing depth or end-effector speed needs to be optimized. For example, during surface sampling, a positional error of 1.2 mm, an orientation error of 2.5°, a contact force fluctuation of ±0.3 N, and disinfectant concentration detection were not completed. Based on optimization rules, the sampling path smoothness coefficient was adjusted from 0.8 to 0.9, the pressing depth from 5 mm to 5.2 mm, and the dwell time from 2 s to 2.5 s. New adjusted sampling parameters were generated for the next round of sampling to reduce positional deviation, stabilize contact pressure, and ensure disinfection effectiveness.

[0054] By dynamically optimizing the sampling standard parameters through real-time data feedback, the robotic sampling process is automated and intelligently adjusted. This not only improves sampling accuracy and stability but also ensures that the sampling tools are disinfected according to preset standards. This guarantees the efficient and reliable operation of the entire sampling and disinfection process, improves the accuracy and reliability of sampling, and ultimately ensures the accuracy and reliability of test results.

[0055] Example 2, based on the same inventive concept as the control and optimization method for the all-media intelligent sampling robot in the foregoing examples, such as... Figure 2 As shown, this application provides a control and optimization system for an all-media intelligent sampling robot, wherein the control and optimization system for the all-media intelligent sampling robot includes: The medium type parsing component 11 is used to receive sampling task instructions, parse the sampling medium type, control the multi-media adaptive end effector to rotate and switch to the corresponding sampling head position and lock it according to the target medium type; the sampling path generation component 12 is used to construct a virtual specification plate space model through infrared ranging and fluorescence projection, perform three-dimensional modeling and boundary calibration of the target sampling area, and generate a standardized sampling path according to the sampling standard parameters in the sampling task instructions; the data acquisition component 13 is used to control the multi-joint robotic arm to perform sampling actions according to the standardized sampling path, and simultaneously acquire the posture data and contact force feedback data of the robotic arm in real time, and execute an in-situ self-disinfection process before and after sampling; the concentration detection component 14 is used to detect the disinfectant concentration in the self-disinfection process and generate disinfection effect verification results; the sampling parameter adjustment component 15 is used to adaptively optimize and adjust the sampling standard parameters based on the posture data, contact force feedback data and disinfection effect verification results to obtain adjusted sampling parameters.

[0056] Furthermore, the medium type resolution component 11 is also used to: the sampling medium type includes air medium, object surface medium and water medium; after obtaining the target medium type, control the multi-medium adaptive end effector to rotate and switch to the corresponding plate holder, swab holder or water sample collection needle module.

[0057] Furthermore, the sampling path generation component 12 is also used to: control the infrared ranging fluorescence projection module to project an array of structured light spots onto the target sampling area, and acquire the distorted image of the structured light spots in physical space through a vision sensor set at the end of the robotic arm; execute a three-dimensional reconstruction algorithm based on the distorted image to obtain spatial point cloud data of the target sampling area, construct a virtual specification board spatial model and complete the boundary calibration of the sampling area; and map the logical point sequence of the virtual specification board to the spatial model according to the sampling standard parameters in the sampling task instruction to generate a standardized sampling path containing spatial position and attitude information.

[0058] Furthermore, the sampling path generation component 12 is also used to: dynamically parameterize the virtual specification board according to the sampling standard parameters in the sampling task instruction to obtain the adjusted virtual specification board; control the end effector of the robotic arm to carry the calibration board to traverse the corner positions of the adjusted virtual specification board, and synchronously collect the actual spatial coordinates of the calibration board by the vision sensor to obtain a logical point sequence; solve the rigid transformation matrix from the projected coordinate system to the robotic arm base coordinate system online using the least squares method; substitute the logical point sequence into the rigid transformation matrix to generate a target pose sequence, and optimize and sort the target pose sequence based on the shortest path principle to obtain a standardized sampling path.

[0059] Furthermore, the concentration detection component 14 is also used to: detect the disinfectant concentration according to a preset detection frequency to obtain a disinfectant concentration sequence; determine whether the last disinfectant concentration in the disinfectant concentration sequence is less than a preset disinfectant concentration threshold; if so, trigger a concentration trend analysis command; based on the concentration trend analysis command, call a trend analyzer to perform trend analysis on the disinfectant concentration sequence, and use the trend analysis result and incomplete disinfection as the disinfection effect verification result; if not, use completed disinfection as the disinfection effect verification result.

[0060] Furthermore, the sampling parameter adjustment component 15 is also used to: perform trajectory deviation analysis on the pose data to obtain position error and attitude error; perform stability analysis on the contact force feedback data to obtain contact pressure fluctuation parameters; and based on the position error, attitude error, contact pressure fluctuation parameters and disinfection effect verification results, adaptively adjust the sampling standard parameters according to preset optimization rules to obtain adjusted sampling parameters.

[0061] Furthermore, the sampling path generation component 12 is also used for: the sampling standard parameters include sampling point spacing, pressing depth, dwell time and path smoothing coefficient.

[0062] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The control and optimization method and specific examples for the all-media intelligent sampling robot in the foregoing embodiment one are also applicable to the control and optimization system for the all-media intelligent sampling robot in this embodiment. Through the foregoing detailed description of the control and optimization method for the all-media intelligent sampling robot, those skilled in the art can clearly understand the control and optimization system for the all-media intelligent sampling robot in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0063] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0064] Obviously, those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for regulating and optimizing an all-media intelligent sampling robot, characterized in that, The method includes: Receive sampling task instructions, parse the sampling medium type, and control the multi-media adaptive end effector to rotate and switch to the corresponding sampling head position and lock it according to the target medium type; A virtual specification board space model is constructed by infrared ranging and fluorescence projection. The target sampling area is then three-dimensionally modeled and its boundaries are calibrated. A standardized sampling path is generated according to the sampling standard parameters in the sampling task instruction. The multi-joint robotic arm is controlled to perform sampling actions according to the standardized sampling path, while the position data and contact force feedback data of the robotic arm are collected in real time, and an in-situ self-disinfection process is performed before and after sampling. The concentration of disinfectant in the self-disinfection process is detected to generate disinfection effect verification results; Based on the pose data, contact force feedback data, and disinfection effect verification results, the sampling standard parameters are adaptively optimized and adjusted to obtain the adjusted sampling parameters. A virtual specification plate spatial model is constructed using infrared ranging and fluorescence projection. The target sampling area is then 3D modeled and its boundaries calibrated. A standardized sampling path is generated based on the sampling standard parameters in the sampling task instruction, including: The infrared ranging fluorescence projection module is controlled to project an array of structured light spots onto the target sampling area, and the distortion image of the structured light spots in physical space is acquired by a vision sensor set at the end of the robotic arm. Based on the distorted image, a three-dimensional reconstruction algorithm is executed to obtain spatial point cloud data of the target sampling area, construct a virtual specification board spatial model, and complete the boundary calibration of the sampling area. According to the sampling standard parameters in the sampling task instruction, the logical point sequence of the virtual specification board is mapped to the spatial model to generate a standardized sampling path containing spatial position and attitude information. Based on the sampling standard parameters in the sampling task instruction, the logical point sequence of the virtual specification board is mapped to the spatial model to generate a standardized sampling path containing spatial position and attitude information, including: The virtual specification board is dynamically parameterized and adjusted according to the sampling standard parameters in the sampling task instruction to obtain the adjusted virtual specification board. The control robot arm carries a calibration board to traverse the corner positions of the adjusted virtual specification board, and the vision sensor synchronously collects the actual spatial coordinates of the calibration board to obtain the logical point sequence. The rigid transformation matrix from the projected coordinate system to the robot arm's base coordinate system is solved online using the least squares method. The logical point sequence is substituted into the rigid transformation matrix to generate the target pose sequence, and the target pose sequence is optimized and sorted based on the shortest path principle to obtain a standardized sampling path.

2. The control and optimization method for an all-media intelligent sampling robot as described in claim 1, characterized in that, The sampling medium types include air, object surface, and water. After obtaining the target medium type, control the multi-media adaptive end effector to rotate and switch to the corresponding plate holder, swab holder, or water sample collection needle module.

3. The control and optimization method for an all-media intelligent sampling robot as described in claim 1, characterized in that, The concentration of disinfectant in the self-disinfection process is detected to generate disinfection effect verification results, including: The concentration of the disinfectant is detected according to a preset detection frequency to obtain a disinfectant concentration sequence; Determine whether the concentration of the last disinfectant in the disinfectant concentration sequence is less than the preset disinfectant concentration threshold. If so, trigger the concentration trend analysis command. Based on the concentration trend analysis command, the trend analyzer is invoked to perform trend analysis on the disinfectant concentration sequence, and the trend analysis results and the incomplete disinfection are used as the disinfection effect verification results. If not, then the completion of disinfection will be taken as the result of disinfection effectiveness verification.

4. The control and optimization method for an all-media intelligent sampling robot as described in claim 1, characterized in that, Based on the pose data, contact force feedback data, and disinfection effect verification results, the sampling standard parameters are adaptively optimized and adjusted to obtain the adjusted sampling parameters, including: Trajectory deviation analysis is performed on the pose data to obtain position error and attitude error; Stability analysis is performed on the contact force feedback data to obtain contact pressure fluctuation parameters; Based on the position error, posture error, contact pressure fluctuation parameters, and disinfection effect verification results, the sampling standard parameters are adaptively adjusted according to the preset optimization rules to obtain the adjusted sampling parameters.

5. The control and optimization method for an all-media intelligent sampling robot as described in claim 4, characterized in that, The sampling standard parameters include the sampling point spacing, pressing depth, dwell time, and path smoothing coefficient.

6. A control and optimization system for an all-media intelligent sampling robot, characterized in that, The step of implementing the control and optimization method for the all-media intelligent sampling robot according to any one of claims 1 to 5, wherein the control and optimization system for the all-media intelligent sampling robot comprises: The media type parsing component is used to receive sampling task instructions, parse the sampling media type, and control the multi-media adaptive end effector to rotate and switch to the corresponding sampling head position and lock it according to the target media type; The sampling path generation component is used to construct a virtual specification board space model through infrared ranging and fluorescence projection, perform three-dimensional modeling and boundary calibration of the target sampling area, and generate a standardized sampling path according to the sampling standard parameters in the sampling task instruction. The data acquisition component is used to control the multi-joint robotic arm to perform sampling actions according to the standardized sampling path, while simultaneously acquiring the robotic arm's pose data and contact force feedback data in real time, and performing an in-situ self-disinfection process before and after sampling. A concentration detection component is used to detect the concentration of disinfectant in the self-disinfection process and generate disinfection effect verification results. The sampling parameter adjustment component is used to adaptively optimize and adjust the sampling standard parameters based on the pose data, contact force feedback data, and disinfection effect verification results to obtain the adjusted sampling parameters.

Images