Intelligent laboratory sample plate transfer control method, system and related device

Abstract

The application provides a sample plate transfer control method, system and related device of an intelligent laboratory. The method comprises the following steps: controlling a humanoid robot to move to a first target position based on initialization task information; determining a grabbing error compensation value based on first sensing data from an environment perception and navigation system; controlling the humanoid robot to perform a material grabbing operation based on the grabbing error compensation value, and performing fine adjustment based on a first feedback signal so that the humanoid robot completes the material grabbing operation; detecting the material grabbing state, and if the state is a completed grabbing state, controlling the humanoid robot to move to a second target position and performing state dynamic adjustment; and placing a target sample plate based on third feedback information when it is detected that the humanoid robot reaches the second target position. In this way, the efficiency and adaptability of the mechanical arm in laboratory automation are improved through scenario-based compliant control and visual-haptic collaborative control, and the sample transfer reliability and efficiency in a high-throughput experiment scene are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and in particular to a sample plate transfer control method, system and related device for an intelligent laboratory. Background Technology

[0002] Intelligent laboratories are demanding increasingly higher levels of automation. Experimental tasks often involve various types, functions, and models of equipment, necessitating the use of robotic arms to move materials between different devices. For example, robotic arms might be used to move sample plates. Many current experimental devices have multiple sample plate storage locations, and each plate station typically has multiple positions. During a single transport operation, a humanoid robot might need to retrieve a sample plate from its starting position at the first plate station of one device and then transfer it to its target position at the second plate station of another device.

[0003] However, current sample transfer schemes based on robotic arms suffer from several drawbacks. The experimental system setup is not streamlined enough, the precision of grasping the sample plate is not high enough, and the sample transfer task is constrained by the robotic arm's picking and moving capabilities, resulting in low efficiency, poor adaptability, and poor reliability. Summary of the Invention

[0004] This application provides a sample plate transfer control method, system, and related device for an intelligent laboratory, which can improve the adaptability of sample plate transfer and enhance the reliability and efficiency of sample transfer in high-throughput experimental scenarios.

[0005] In a first aspect, embodiments of this application provide a sample plate transfer control method for an intelligent laboratory, applied to a controller of a humanoid robot system. The humanoid robot system further includes the humanoid robot, which includes a robotic arm equipped with a gripping device for grasping materials, the materials including a target sample plate. The humanoid robot system also includes a plate station module and an environmental perception and navigation system. The plate station module includes a first target position and a second target position. The method includes:

[0006] Acquire task information and determine initialization task information based on the task information, and control the humanoid robot to move from the initial position to the first target position based on the initialization task information;

[0007] The first sensing data of the humanoid robot at the first target position is obtained from the environmental perception and navigation system, and the grasping error compensation value is determined based on the first sensing data. The first sensing data includes first visual data and first mechanical data.

[0008] Based on the grasping error compensation value, the humanoid robot is controlled to perform a material grasping operation on the target sample plate, and a first feedback signal is obtained. Based on the first feedback signal, the humanoid robot is fine-tuned to complete the material grasping operation. The first feedback signal is the second visual data and / or the second mechanical data detected during the material grasping operation.

[0009] The material grasping status is detected. If the material grasping status is detected as completed, the humanoid robot is controlled to move to the second target position based on the initialization task information. During the movement, the status is dynamically adjusted based on the second feedback information. The second feedback information includes the third visual data and / or the third mechanical data detected during the movement.

[0010] When the humanoid robot is detected to have reached the second target position, third feedback information is obtained and the target sample plate is placed based on the third feedback information. The third feedback information includes fourth visual data and / or fourth mechanical data.

[0011] In a second aspect, embodiments of this application provide a humanoid robot system that applies the sample plate transfer control method of the intelligent laboratory as described in the first aspect, comprising:

[0012] A humanoid robot, comprising a mobile chassis and a robotic arm, wherein the robotic arm is equipped with a gripping device for gripping materials, the materials including the target sample plate, and the mobile chassis comprising a bipedal structure and / or a wheeled structure, wherein the humanoid robot switches to the wheeled structure when moving and switches to the bipedal structure when standing.

[0013] A controller for performing the method described in the first aspect above;

[0014] An environmental perception and navigation system, comprising a vision system, a six-dimensional force sensor, and a navigation and positioning system, wherein the vision system is used to acquire visual data, the six-dimensional force sensor is used to acquire mechanical data, and the navigation and positioning system is used to provide path information and the intelligent laboratory space information;

[0015] The board station module is a multi-modal area for storing sample boards, and the board station module is a multi-modal design.

[0016] In one possible embodiment, the vision system includes a plurality of camera devices, including a head-mounted camera device and a robotic arm camera device;

[0017] The six-dimensional force sensor includes a robotic arm force sensor for sensing the force on the gripping device and / or the end effector of the robotic arm; the plate station is a multi-layer plate station built with a C-shaped channel steel frame, and the plate station includes an operating area, with at least one positioning code set in each operating area.

[0018] The controller, the head camera device, the robotic arm camera device, and the six-dimensional force sensor are synchronized in time and space;

[0019] The humanoid robot's robotic arm includes a robotic arm with 7 degrees of freedom.

[0020] Thirdly, embodiments of this application provide a sample plate transfer control device for an intelligent laboratory, applied to a controller of a humanoid robot system. The humanoid robot system further includes the humanoid robot, which includes a robotic arm equipped with a gripping device for grasping materials, including target sample plates. The humanoid robot system also includes a plate station module and an environmental perception and navigation system. The plate station module includes a first target position and a second target position. The device includes:

[0021] An initialization module is used to acquire task information and determine initialization task information based on the task information, and to control the humanoid robot to move from an initial position to the first target position based on the initialization task information;

[0022] The first material grasping module is used to acquire first sensing data of the humanoid robot at the first target position from the environmental perception and navigation system, and determine the grasping error compensation value based on the first sensing data. The first sensing data includes first visual data and first mechanical data.

[0023] The second material grasping module is used to control the humanoid robot to perform material grasping operation on the target sample plate based on the grasping error compensation value, and to obtain a first feedback signal. Based on the first feedback signal, the module makes fine adjustments to enable the humanoid robot to complete the material grasping operation. The first feedback signal is second visual data and / or second mechanical data detected during the material grasping operation.

[0024] The material transfer module is used to detect the material grasping status. If the material grasping status is detected as completed, the module controls the humanoid robot to move to the second target position based on the initialization task information. During the movement, the module dynamically adjusts the status based on the second feedback information. The second feedback information includes the third visual data and / or the third mechanical data detected during the movement.

[0025] The material placement module is used to obtain third feedback information and place the target sample plate based on the third feedback information when the humanoid robot is detected to have reached the second target position. The third feedback information includes fourth visual data and / or fourth mechanical data.

[0026] Fourthly, embodiments of this application provide an electronic device, including a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the programs include instructions for performing some or all of the steps described in the first aspect of embodiments of this application.

[0027] Fifthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, the computer program including program instructions that, when executed by a processor, cause the processor to perform some or all of the steps described in the first aspect.

[0028] Sixthly, embodiments of this application provide a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps described in the first aspect of embodiments of this application. The computer program product may be a software installation package.

[0029] By implementing the embodiments of this application, task information is acquired and initialization task information is determined based on the task information; and the humanoid robot is controlled to move from an initial position to a first target position based on the initialization task information; first sensing data of the humanoid robot at the first target position is acquired from the environmental perception and navigation system, and a grasping error compensation value is determined based on the first sensing data, wherein the first sensing data includes first visual data and first mechanical data; the humanoid robot is controlled to perform a material grasping operation on the target sample plate based on the grasping error compensation value, and a first feedback signal is acquired; fine-tuning is performed based on the first feedback signal to enable the humanoid robot to complete the material grasping operation. The first feedback signal is the second visual data and / or second mechanical data detected during the material grasping operation. The material grasping state is detected; if the grasping state is detected as completed, the humanoid robot is controlled to move to the second target position based on the initialization task information. During the movement, the state is dynamically adjusted based on the second feedback information, which includes the third visual data and / or third mechanical data detected during the movement. When the humanoid robot reaches the second target position, the third feedback information is acquired, and the target sample plate is placed based on the third feedback information, which includes the fourth visual data and / or fourth mechanical data. Thus, through scenario-based compliant control and vision-force coordinated control, the efficiency and adaptability of the robotic arm in laboratory automation are improved, significantly enhancing the reliability and efficiency of sample transfer in high-throughput experimental scenarios. Attached Figure Description

[0030] To more clearly illustrate the technical solutions in the embodiments of the present invention or the background art, the accompanying drawings used in the embodiments of the present invention or the background art will be described below.

[0031] Figure 1a This is a schematic diagram of the architecture of a humanoid robot system provided in an embodiment of this application;

[0032] Figure 1b This is a schematic diagram of the architecture of another humanoid robot system provided in the embodiments of this application;

[0033] Figure 1c This is a structural schematic diagram of a humanoid robot provided in an embodiment of this application;

[0034] Figure 2 This is a schematic flowchart of a sample plate transfer control method for an intelligent laboratory provided in an embodiment of this application;

[0035] Figure 3 This is a schematic flowchart of another intelligent laboratory sample plate transfer control method provided in this application embodiment;

[0036] Figure 4 This is a schematic diagram of the structure of a sample plate transfer control device for an intelligent laboratory proposed in an embodiment of this application;

[0037] Figure 5 This is a schematic diagram of the structure of another intelligent laboratory sample plate transfer control device provided in the embodiments of this application;

[0038] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0039] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application.

[0040] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or electronic device that includes a series of steps or units is not limited to the listed steps or units, but in an alternative example also includes steps or units not listed, or in an alternative example also includes other steps or units inherent to these processes, methods, products, or electronic devices.

[0041] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0042] Current sample transfer schemes based on robotic arms suffer from several drawbacks: the experimental system setup is not streamlined enough, the precision of grasping the sample plates is not high enough, and the sample transfer task is constrained by the robotic arm's picking and moving capabilities, resulting in low efficiency, poor adaptability, and poor reliability.

[0043] To address the aforementioned issues, this application provides a sample plate transfer control method, system, and related device for an intelligent laboratory. Through scenario-based compliant control and vision-force coordinated control, the efficiency and adaptability of the robotic arm in laboratory automation are improved, significantly enhancing the reliability and efficiency of sample transfer in high-throughput experimental scenarios.

[0044] The sample plate transfer control method for intelligent laboratories provided in this application embodiment can be applied to, for example... Figure 1a Please refer to the humanoid robot system shown. Figure 1a and Figure 1b , Figure 1a This is a schematic diagram of the architecture of a humanoid robot system provided in an embodiment of this application. Figure 1b This is a schematic diagram of the architecture of another humanoid robot system provided in an embodiment of this application; as shown Figure 1a As shown, the humanoid robot system 100 includes a humanoid robot 110, a controller 120, an environmental perception and navigation system 130, and a board module 140. The humanoid robot 110 is controlled by the controller 120. Figure 1b As shown, the humanoid robot 110, controller 120, and environmental perception and navigation system 130 can be integrated into one unit.

[0045] In this scheme, controller 120 refers to a computer used to handle a large number of computational tasks and store data. In this scheme, controller 120 is equipped with various models for executing the sample plate transfer control method in the intelligent laboratory. Controller 120 can also be used to collect data during model usage, facilitating subsequent optimization. The environmental perception and navigation system 130, through multi-sensor collaboration, provides the humanoid robot 110 with environmental information perception and path planning capabilities. The environmental perception and navigation system 130 includes a vision system, a six-dimensional force sensor, and a navigation and positioning system. The vision system is used to acquire visual data, the six-dimensional force sensor is used to acquire mechanical data, and the navigation and positioning system is used to provide path information and the spatial information of the intelligent laboratory. The plate station module 140 consists of multiple modular areas for storing sample plates, and the plate station module 140 is a multi-modal design. Please refer to [link to relevant documentation]. Figure 1c , Figure 1cThis is a schematic diagram of the structure of a humanoid robot provided in an embodiment of this application. The humanoid robot 110 includes a mobile chassis and a robotic arm. The robotic arm is equipped with a gripping device for gripping materials, including the target sample plate. The mobile chassis includes a bipedal structure and / or a wheeled structure. When the humanoid robot 110 moves, it switches to the wheeled structure; when the humanoid robot 110 stands, it switches to the bipedal structure. Specifically, when the humanoid robot 110 moves from an initial position to a first target position, from the first target position to a second target position, and from the second target position to other positions, the humanoid robot 110 moves using the wheeled structure. When the humanoid robot 110 pauses at the first target position or the second target position, the humanoid robot 110 stands using the bipedal structure. The specific size parameters of the humanoid robot 110, such as its width and height, are related to the spatial constraints of key areas in the intelligent laboratory. For example, if the door width of the intelligent laboratory is 80 cm, then the width of the humanoid robot 110 should be adjusted to be less than or equal to 75 cm. The humanoid robot 110 can also be equipped with an anti-collision device. The contact force threshold of this device can be set according to requirements, for example, the contact force can be greater than or equal to 20N. The aforementioned gripping device can be an adaptive electric gripper (supporting automatic recognition of 96-hole plates / 384-hole plates), and the gripping force range can be within a preset force range, such as 0.5-5N (adjustable). The type of gripper can be changed according to different application requirements or scenarios, specifically through magnetic attraction, knobs, etc. The aforementioned robotic arm can have two, namely a dual-arm robotic arm, with 7 degrees of freedom, including 3 degrees of freedom in the shoulder, 2 degrees of freedom in the elbow, and 2 degrees of freedom in the wrist. The load capacity of the robotic arm can be 3kg (continuous) / 5kg (peak) per arm. The working range of the robotic arm is a combined arm extension radius of 1.8m. The humanoid robot 110 has a built-in 20Ah lithium battery, with a battery life of 8 hours. The humanoid robot 110 also includes a head, and the robotic arm is equipped with a robotic arm camera device. Specifically, the vision system mainly consists of at least one head camera device installed on the head of the humanoid robot 110 and at least one robotic arm camera device at the end of the robotic arm. The head camera device can be an RGB-D camera (resolution up to 1920×1080). The RGB-D camera on the head can acquire large-area color images and depth information in real time, used to construct 3D maps of intelligent laboratories, identify the location of laboratory equipment, and detect obstacles. The robotic arm camera device at the end of the robotic arm is used for close-range high-definition imaging and depth perception of target areas. In some cases, the vision system also includes 3D structured light, which can work in conjunction with the robotic arm camera device to achieve close-range depth perception, with an effective range of 0.5-5m.

[0046] Specifically, a six-dimensional force sensor is installed at the end effector of the robotic arm, with a force resolution of 0.1N. It is used to detect the forces acting on the humanoid robot 110 end effector in three force directions (X, Y, and Z axes) and three torque directions (around the X, Y, and Z axes) when it comes into contact with an external object. During the grasping of the target sample plate, the six-dimensional force sensor data is fused with the vision system to compensate for vibration errors of the humanoid robot 110, ensuring accurate grasping by the gripping device. During contact operations, a force control mode is implemented based on force sensor feedback, allowing the gripping device to adjust the clamping force according to the sample plate material, ensuring operational safety and stability.

[0047] Specifically, the navigation and positioning system employs simultaneous localization and mapping (SLAM) technology. During the task initialization phase, the humanoid robot 110 uses a head-mounted 3D camera and LiDAR in collaboration to scan the work area, constructing a 3D map in real time that includes information such as the board station's location, obstacles, and laboratory equipment. It then determines its own position on the map and assesses obstacle conditions in real time. The navigation and positioning system has an accuracy of ±10mm, meeting the positioning requirements of the humanoid robot 110 for large-scale movement within the laboratory. Combining path planning algorithms such as the A* algorithm and the fast randomized tree search algorithm, the humanoid robot 110 can achieve optimal path planning from its current position to the target position. Furthermore, during movement, a dynamic window algorithm is used to dynamically adjust the path based on environmental changes.

[0048] Specifically, the board station module 140 is a multi-layer board station constructed with a C-shaped channel steel frame. The height of each layer is adjustable, specifically between 50-300mm. The board station module includes an operating area, and each operating area is equipped with at least one positioning code. This positioning code can be a sticker affixed to the operating area, and the spacing between each QR code can be 50cm. The board station module is equipped with light strips to indicate the board position status, such as whether the space is empty. In the board station module 140, landmarks are set on the critical path. These landmarks can be reflective, luminous, or have a specific color for use in conjunction with the LiDAR for navigation.

[0049] Based on this, this application provides a sample plate transfer control method for an intelligent laboratory, which will be described in detail below with reference to the accompanying drawings.

[0050] Please see Figure 2 , Figure 2This is a flowchart illustrating a sample plate transfer control method for an intelligent laboratory according to an embodiment of this application. The method is applied to a controller of a humanoid robot system. The humanoid robot system further includes the humanoid robot itself, which includes a robotic arm equipped with a gripping device for grasping materials, including a target sample plate. The humanoid robot system also includes a plate station module and an environmental perception and navigation system. The plate station module includes a first target position and a second target position. Figure 2 As shown, the method includes the following steps:

[0051] S210, acquire task information and determine initialization task information based on the task information, and control the humanoid robot to move from the initial position to the first target position based on the initialization task information.

[0052] The host computer communicates with the humanoid robot system. When a laboratory operator initiates a sample plate transfer task through the host computer, the host computer sends a task command to the controller via Modbus TCP / IP and ROS2 standard interfaces. This task command includes task information such as the station ID of the first target location, which specifies the initial storage location of the target sample plate (e.g., device A - station 1), the station ID of the second target location, which specifies the required placement location of the target sample plate (e.g., device B - station 3), the type of target sample plate (e.g., different specifications such as 96-well / 384-well, used for subsequent adaptive control of the gripping device and selection of operating strategies), and priority (e.g., high, medium, low levels), providing a basis for task scheduling optimization, with high-priority tasks being executed first.

[0053] Upon receiving the task instruction, the robot immediately activates the SLAM module during the task initialization phase. At this time, the environmental perception and navigation system is activated, collecting environmental images and depth data, specifically information such as the board location, laboratory equipment, and obstacles. Simultaneously, it performs high-frequency scanning of the surrounding environment in real time, generating a high-precision point cloud map. Through a data fusion algorithm, the visual image data and LiDAR point cloud data are integrated to construct a 3D map containing information such as the board location, laboratory equipment, and obstacles. After constructing the 3D map, the controller uses the A* algorithm for path planning. The A* algorithm calculates the optimal path from the robot's current position to the source board by comprehensively considering the distance cost of the path and the heuristic function evaluation value, reserving a redundancy space. Specifically, the redundancy space can be 20%, which is not limited here, to cope with possible sudden obstacles or environmental changes.

[0054] In some possible scenarios, when the humanoid robot starts from its initial position, its chassis switches to a wheeled structure, allowing it to move along an optimal path at a preset speed. During this movement, a 16-line LiDAR continuously scans the surrounding environment at a frequency of 20Hz, detecting obstacle information in real time. Once an obstacle is detected, the LiDAR immediately transmits the data to the robot's control system, triggering a dynamic window approach (DWA) algorithm for local path adjustment. Based on the robot's current speed, acceleration limits, and obstacle distribution, the DWA algorithm quickly replans a safe and feasible local path (response time < 150ms), ensuring the robot can flexibly avoid obstacles and continue moving towards the source board station. When the robot reaches a preset stopping distance directly in front of the first target position, its chassis switches to a bipedal structure to reduce space occupation and adapt to the needs of close-range precision operation. Simultaneously, the humanoid arm is pre-raised to a safe height above the tabletop, preventing collisions with the tabletop or other objects.

[0055] S220, acquire first sensing data of the humanoid robot at the first target position from the environmental perception and navigation system, and determine a grasping error compensation value based on the first sensing data, wherein the first sensing data includes first visual data and first mechanical data.

[0056] When the humanoid robot moves to the first target position, the robotic arm camera at the end of the robotic arm works in conjunction with the head camera to acquire the first visual data. The robotic arm camera focuses on the target plate, capturing high-definition images at a frame rate of at least 30Hz to obtain close-range visual data including details such as the first positioning code and edge features of the target sample plate. The head camera simultaneously acquires depth information and color images of the surrounding environment to help determine the spatial relationship between the robot and the target sample plate. Combined with the YOLOv8 model, the images are processed in real time to identify the first positioning code and output its initial coordinates in the robot's coordinate system. A six-dimensional force sensor monitors the interaction forces between the gripping device and the external environment in real time. During the approach to the target sample plate, it continuously acquires force data in three force directions (X, Y, Z axes) and three torque directions (around the X, Y, Z axes), recording minute force fluctuations caused by robot vibration and joint inertia. Simultaneously, a 9-axis IMU synchronously detects the acceleration and angular velocity information of the robot arm to obtain the first mechanical data. Among them, the compensation force value, i.e. the error compensation value, is determined based on the minute force fluctuations of the body vibration contained in the first mechanical data.

[0057] In one possible embodiment, the first visual data comes from the robotic arm camera device. Acquiring the first sensing data of the humanoid robot at the first target position and determining a grasping error compensation value based on the first sensing data includes: acquiring first platform coordinate information based on a first positioning code, where the first positioning code is the positioning code for the first target position, and the first platform coordinate information indicates the precise coordinates of the platform where the first target position is located; determining a first damping parameter and a first vibration frequency parameter based on the first mechanical data, where the first damping parameter and the first vibration frequency parameter characterize the real-time vibration and damping of the humanoid robot's joints; and determining a first error compensation value based on the first damping parameter and the first platform coordinate information, where the first error compensation value includes a first target damping coefficient, which is used to adjust and suppress errors caused by robotic arm vibration.

[0058] The process involves real-time image detection using the YOLOv8 model to quickly locate the first positioning code region and extract the geometric feature points of the positioning code to obtain the precise coordinates of the first target location on the board. A six-dimensional force sensor acquires force data in real-time at a high-frequency sampling rate of 1000Hz, while a 9-axis IMU synchronously monitors inertial parameters. Vibration and damping characteristics are then analyzed to obtain the first damping parameter and the first error compensation value. Specifically, this involves performing a Fast Fourier Transform (FFT) on the time-domain signal from the six-dimensional force sensor to convert the force data to the frequency domain. Then, by identifying the peak frequencies in the spectrum, the main frequency components of the robot joint vibration are determined, i.e., the first vibration frequency parameter. Combined with the robot's dynamics model, a mathematical relationship between force fluctuation amplitude and joint damping is established. For example, by analyzing the attenuation rate of the force signal within the vibration cycle, the equivalent damping coefficient of each joint is calculated. Simultaneously, an Extended Kalman Filter (EKF) is used to filter the force data, separating the damping force components caused by mechanical friction, transmission clearance, and other factors to obtain the first damping parameter. This parameter reflects the energy loss characteristics of the joint during movement and is used to assess the damping state of the current mechanical structure. Then, through a data fusion algorithm, visual positioning information is combined with mechanical characteristic parameters to generate a vibration compensation strategy, i.e., the first target damping coefficient. Specifically, this may include: establishing a vibration displacement prediction model for the robot's end effector based on the first vibration frequency parameter and the coordinate information of the first plate station; using the inverse kinematics of the robot to map the vibration frequency to the fluctuation range of the end effector position; assessing the current joint's vibration reduction capability based on the first damping parameter, and combining this with the predicted vibration error, using a PID control algorithm to back-calculate the required first target damping coefficient. For example, if insufficient damping leads to excessive vibration error, the target damping coefficient is increased to enhance the joint's damping characteristics. The first target damping coefficient is then converted into control parameters for the joint servo system, generating a first error compensation value that includes position compensation (ΔX, ΔY, ΔZ) and attitude compensation (Δθx, Δθy, Δθz).

[0059] As can be seen, in this embodiment, the vibration generated by the robotic arm during the grasping process is compensated based on the first mechanical data and the first visual data. The synergy between vision and force makes the operation more stable and precise.

[0060] S230, based on the grasping error compensation value, the humanoid robot is controlled to perform a material grasping operation on the target sample plate, and a first feedback signal is obtained. Based on the first feedback signal, fine-tuning is performed to enable the humanoid robot to complete the material grasping operation. The first feedback signal is second visual data and / or second mechanical data detected during the material grasping operation.

[0061] Specifically, the grasping error compensation value calculated above compensates for the vibration generated during the extension of the humanoid robot's robotic arm in real time, and detects the first feedback signal in real time. The first feedback signal is used to determine whether the target sample plate is about to be touched or has already been touched. Specifically, it includes at least one of second visual data and second mechanical data. In this embodiment, the acquisition of second visual data and second mechanical data is taken as an example. The second visual data includes data captured by the robotic arm's camera device, and the second mechanical data is data acquired by a six-dimensional force sensor.

[0062] Specifically, the sample plate contour is identified using an edge detection algorithm and matched with a preset template to calculate the deviation between the actual gripping position and the ideal position. Simultaneously, a deep learning model is used to perform semantic analysis on the image to determine if the sample plate has tilted, shifted, or experienced other anomalies. A six-dimensional force sensor monitors the force data in three force directions (X, Y, and Z axes) and three torque directions (around the X, Y, and Z axes) in real time. By analyzing the time-domain changes in the force signals, it determines whether the gripping device is stably gripping the sample plate. For example, if the force value in a certain direction exhibits periodic fluctuations and the amplitude exceeds a threshold (e.g., ±0.2N), it indicates vibration or contact instability, generating a mechanical feedback signal. Furthermore, a 9-axis IMU detects the acceleration and angular velocity information of the robotic arm to assist in determining sample plate swaying caused by the robot's movement.

[0063] Specifically, if the visual feedback signal indicates an angular deviation in the sample plate, the feedback control module calculates the required joint angle adjustment value based on the deviation and drives the wrist joint (2 degrees of freedom) to perform fine-tuning rotation through a PID control algorithm, achieving an adjustment accuracy of ±0.1°. Simultaneously, it combines torque data from a six-dimensional force sensor to ensure stable gripping force during adjustment. When the mechanical feedback signal indicates uneven gripping force or vibration, the system first uses 9-axis IMU data to determine the source of vibration (e.g., chassis movement, robotic arm resonance). If the vibration is from the robotic arm itself, the feedback control module adjusts the damping parameters and driving torque of each joint based on the vibration frequency and amplitude to suppress the vibration. If the uneven force is caused by positional deviation, it calculates the position compensation based on visual data and drives the robotic arm to perform small-amplitude translational adjustments (accuracy ≤0.1mm).

[0064] In one possible embodiment, the second visual data in the first feedback signal comes from the robotic arm camera device, and the initialization task information further includes a material type, which is used to indicate the characteristics of the target sample plate. Acquiring the first feedback signal and fine-tuning based on the first feedback signal to complete the material gripping operation includes: continuously acquiring the first feedback signal during the gripping process; if the first feedback signal is detected to meet a first contact condition, adjusting the touch force based on the material type, where the first contact condition is used to constrain the relative distance and / or contact force between the humanoid robot and the target sample plate; determining a first motor torque based on the material type and the touch force, and completing the material gripping operation based on the first motor torque; determining a current center of gravity parameter based on the first feedback signal, where the center of gravity parameter characterizes the center of gravity of the target sample plate relative to the humanoid robot's grip; determining a second motor torque based on the center of gravity parameter, and continuously detecting the center of gravity offset, where the center of gravity offset characterizes the difference between the center of gravity parameter and the target center of gravity parameter; and adjusting the center of gravity based on the center of gravity offset to smoothly complete the material gripping operation.

[0065] The initial task information may include the material, style, and size of the target sample plate, which determines the choice of gripping strategy. During the gripping process, the first feedback signal (second visual data and second mechanical data) is monitored in real time. When the first feedback signal meets the first contact condition, a targeted touch force adjustment mechanism is triggered. This first contact condition can be defined as the relative distance between the gripping device and the sample plate decreasing to below 0.1 cm, and the six-dimensional force sensor detecting a normal contact force greater than or equal to 0 N. The mapped touch force is determined based on the material type; different material types are mapped to corresponding touch forces, for example, 1.5 N for plastic plates and 2.5 N for glass plates. Based on the robot's dynamics model and the mechanical structure parameters of the gripping device, a mathematical relationship is established between motor torque, touch force, and material characteristics. The formula is as follows:

[0066] T = k × F grip ×r×μ;

[0067] Where T is the torque of the first motor, k is the mechanical transmission coefficient, and F gripThe force applied is denoted by 'r', the distance from the gripping point to the motor shaft, and 'μ', the coefficient of friction of the material surface, which can be determined by the material type. This model calculates the motor torque required for stable gripping, ensuring the target sample plate is firmly held. Then, the center of gravity parameters of the sample plate are determined using IMU and force sensor data, and dynamic balance adjustments are performed. Specifically, a 9-axis IMU monitors the acceleration and angular velocity of the robot arm in real time, combined with force distribution data detected by a six-dimensional force sensor, to calculate the center of gravity position of the sample plate relative to the gripping device using mechanical balance equations. The calculated center of gravity parameters are compared with the target center of gravity parameters (the center of gravity position under ideal equilibrium conditions) to calculate the center of gravity offset. When the offset exceeds a threshold (e.g., 2mm), the required angle adjustment for each joint is calculated using an inverse kinematics algorithm, thus determining the second motor torque. The control system drives the robot arm joints to adjust the arm posture, bringing the center of gravity back to the target range. During this process, the system continuously monitors the center of gravity offset at a frequency of no less than 100Hz, achieving dynamic closed-loop adjustment through PID control.

[0068] As can be seen, in this embodiment, the grasping control process based on material type and real-time feedback combines mechanical and visual data, allowing for slight compliant movements during contact to avoid hard collisions. The robot simultaneously performs plate position recognition and pre-adjustment of the gripping device during movement, which improves the efficiency and adaptability of the robotic arm in laboratory automation and significantly enhances the reliability and efficiency of sample transfer in high-throughput experimental scenarios.

[0069] S240, detect the material grasping state. If the material grasping state is detected as completed, control the humanoid robot to move to the second target position based on the initialization task information, and dynamically adjust the state based on the second feedback information during the movement. The second feedback information includes the third visual data and / or the third mechanical data detected during the movement.

[0070] The material grasping status can be determined using at least one of visual and mechanical methods. Visual data can be continuously received images from the robotic arm's camera device, while mechanical data can be obtained by continuously monitoring the contact force between the gripper and the target sample plate using a six-dimensional force sensor. The material grasping status can be determined using one or both of these visual and mechanical data. The material grasping status includes "grab completed." After confirming that the grasping status is complete, the robot performs a movement operation based on the target plate station ID (i.e., the second target position) in the initial task information, enabling the target sample plate to be transferred smoothly. During the movement, the robot collects third visual data and third mechanical data in real time as second feedback information for dynamic status adjustment. Specifically, the third visual data can be collected by at least one of the head-mounted camera device and the robotic arm's camera device, and the third mechanical data can be collected by the six-dimensional force sensor.

[0071] Specifically, the head-mounted camera captures environmental images at a frequency of at least 10Hz, and combined with LiDAR data, uses a semantic segmentation algorithm to identify obstacle types (such as static equipment and dynamic personnel) and their locations in real time. When an obstacle is detected approaching a preset safe distance, the robot adjusts its speed and direction according to the DWA algorithm to achieve dynamic obstacle avoidance. The robotic arm's camera continuously monitors the levelness of the sample plate, detecting its posture changes at a frequency of at least 10Hz. When the third-vision data shows that the levelness deviation of the sample plate exceeds 0.5°, the system immediately triggers a micro-adjustment mechanism, controlling the joint angles of the robotic arm to restore the sample plate's posture to the allowable range, with an adjustment accuracy of ±0.1°. A six-dimensional force sensor and a 9-axis IMU monitor the force and vibration status of the robotic arm in real time. When the third-dimensional force data shows abnormal force fluctuations caused by chassis movement or robotic arm resonance, the system adjusts the damping parameters and driving torque of each joint according to the vibration frequency and amplitude using a PID algorithm to suppress vibration. The six-dimensional force sensor continuously monitors the contact force between the gripper and the sample plate. If the contact force deviates from the target touch force by more than the allowable range (e.g., ±0.2N), the system determines that there may be a risk of the sample plate becoming loose, and immediately adjusts the motor torque of the gripping device to restore the contact force to a stable state, ensuring that the sample plate does not fall off during movement.

[0072] In one possible embodiment, the third visual data comes from the head camera device and / or the robotic arm camera device. The dynamic state adjustment based on the second feedback information during movement includes: receiving the second feedback information from the environmental perception and navigation system; when a sudden obstacle is detected based on the second feedback information, controlling the humanoid robot to brake suddenly and replanning the path based on the environmental perception and navigation system; determining chassis vibration compensation data based on a PID algorithm, determining levelness data based on the second feedback information, wherein the chassis vibration compensation data is used to reduce the degree of chassis vibration of the humanoid robot, and the levelness data is the levelness data of the target sample plate; and performing the dynamic state adjustment of the humanoid robot based on the chassis vibration compensation data and the levelness data.

[0073] The system comprises several components: a head-mounted camera that acquires environmental images at a frequency of at least 10Hz, combined with LiDAR data, and uses a semantic segmentation algorithm to identify obstacle types (such as static equipment and dynamic personnel) and their locations in real time; a robotic arm camera that continuously monitors the levelness of the sample plate and detects its posture changes at a frequency of at least 10Hz; and third-order visual data based on these parameters. A six-dimensional force sensor and a nine-axis IMU monitor the force and vibration of the robotic arm in real time, generating third-order mechanical data. The control system employs a deep learning-based multimodal fusion algorithm to obtain second-order feedback information, spatially aligns the LiDAR point cloud with the visual images, uses semantic segmentation technology to identify obstacle categories (static equipment, dynamic personnel, etc.), and calculates parameters such as distance and speed relative to the robot. For sample plate levelness detection, the robotic arm camera uses edge detection and posture estimation algorithms to compare real-time pose data with the ideal level state, outputting levelness data with an accuracy of 0.1°.

[0074] The LiDAR and vision system assess obstacle risk using preset thresholds. For example, when an obstacle is detected to be less than a safe threshold (e.g., 0.5m) and the relative distance exceeds a warning value, an emergency braking command is triggered. The control system quickly cuts off power to the wheel drive motors and simultaneously activates the electromagnetic braking device, controlling the robot's braking distance to within 0.3m (in wheeled mode) to avoid collisions. After braking, the robot uses the RRT* algorithm for path replanning based on a real-time map built using SLAM. This algorithm uses the current position as the starting point and the target platform as the ending point, optimizing path length and smoothness while avoiding obstacles. During this process, a 9-axis IMU monitors the chassis vibration frequency and amplitude in real time, and a six-dimensional force sensor captures the vibration stress transmitted by the robotic arm. Compensation parameters are calculated based on a PID algorithm. For example, when a low-frequency vibration of 15Hz is detected, the PID controller outputs a compensation signal to the joint actuator, dynamically adjusting the motor damping coefficient and driving torque to suppress vibration transmission. Based on the levelness data acquired by the robotic arm's camera device, the system establishes a closed-loop control circuit. When the level deviation exceeds a threshold (e.g., 0.5°), the control system calculates the adjustment amount of the robotic arm joints using an inverse kinematics algorithm, prioritizing the driving of the two degrees of freedom of the wrist for fine-tuning of the posture, with an adjustment accuracy of ±0.1°. The controller integrates the path replanning results, vibration compensation parameters, and level adjustment commands, and synchronously distributes them to each execution module of the robot through the service interface of the ROS2 framework (e.g., / motion_control).

[0075] As can be seen, in this embodiment, through high-frequency data acquisition from lidar and visual sensors, and multimodal fusion algorithms, the system can detect dynamic obstacles in real time. The chassis vibration compensation mechanism based on the PID algorithm significantly reduces end-effector displacement errors caused by vibration. Combined with a six-dimensional force sensor for real-time monitoring of gripping force fluctuations (accuracy up to ±0.1N), sample plate damage is effectively avoided. Based on visual feedback and inverse kinematics algorithms from the robotic arm's camera device, level deviation correction can be quickly completed, ensuring that the ideal posture is maintained during transfer. Thus, the efficiency and adaptability of the robotic arm in laboratory automation are improved, significantly enhancing the reliability and efficiency of sample transfer in high-throughput experimental scenarios.

[0076] S250, when the humanoid robot is detected to have reached the second target position, a third feedback information is obtained and the target sample plate is placed based on the third feedback information, wherein the third feedback information includes fourth visual data and / or fourth mechanical data.

[0077] The second target position is the target transfer position of the target sample plate. The acquisition methods of the fourth mechanical data and the fourth visual data in the third feedback information are the same as those of the first feedback information, and will not be repeated here.

[0078] In one possible embodiment, the fourth visual data comes from the robotic arm camera device. The step of acquiring third feedback information and placing the target sample plate based on the third feedback information includes: acquiring second plate station coordinate information based on the fourth visual data using a second positioning code, where the second positioning code is the positioning code for the second target position, and the second plate station coordinate information is used to indicate the precise coordinates of the plate station where the first target position is located; acquiring a first empty space status for the target placement position, where the target placement position is the empty space indicated by the initialization task information for placing the target sample plate, and the first empty space status includes an idle state; if the first empty space status is the idle state, controlling the humanoid robot to perform a material placement operation, during the material placement operation, determining a second damping parameter and a second vibration frequency parameter based on the fourth mechanical data, where the second damping parameter and the second vibration frequency parameter are used to characterize the humanoid robot. The system monitors the real-time vibration and damping of the joints; determines a second error compensation value based on the second damping parameter and the second plate coordinate information, the second error compensation value including a second target damping coefficient, which is used to adjust and suppress the error caused by the vibration of the robotic arm; continuously acquires a third feedback signal during placement; if the third feedback signal is detected to meet the second contact condition, adjusts the touch force based on the material type, the second contact condition being used to constrain the relative distance and / or contact force between the humanoid robot and the target placement position; determines the first motor torque based on the material type and the touch force, and completes the material placement operation based on the first motor torque; when it is detected that the target sample plate has detached from the gripping device, controls the robotic arm to lift up and acquires the target placement position image of the robotic arm camera device; determines the placement result based on the target position placement image, the placement result including that the sample plate has been seated.

[0079] The acquisition method for the third feedback information is the same as that for the first feedback information, and will not be repeated here. This yields the fourth visual data and the fourth mechanical data. The fourth visual data also includes the coordinate information of the second board station obtained by scanning the second positioning code, as well as the first empty space status. The first empty space status refers to the empty space status of the board station. When no obstruction is detected at the target board position, it is determined to be in an empty state, and this information is fed back to the robot control system. Only when the first empty space status is empty will the robot control system trigger the material placement operation command; otherwise, it will repeatedly determine the first empty space status or report it. During the material placement operation, the six-dimensional force sensor and 9-axis IMU at the end of the robotic arm collect the fourth mechanical data in real time for analyzing joint vibration and damping. During the material placement operation, the process of determining the second damping parameter and the second vibration frequency parameter based on the fourth mechanical data, and the process of determining the second error compensation value based on the second damping parameter and the second plate station coordinate information, is the same as the process of determining the first damping parameter and the first vibration frequency parameter based on the first mechanical data, and the process of determining the first error compensation value based on the first damping parameter and the first plate station coordinate information, as described above. It will not be repeated here, but you can refer to the detailed description above.

[0080] During placement, the system continuously collects third feedback signals (including fourth visual and fourth mechanical data) to dynamically adjust the placement operation: determining whether the second contact condition is met. For example, when the relative distance between the gripping device and the target placement position drops below 0.1 cm, and the six-dimensional force sensor detects a normal contact force greater than 0 N, the second contact condition is deemed met. Based on the material type in the initial task information, the control system retrieves the corresponding target touch force from a pre-stored parameter table (e.g., 1.5 N for plastic plates and 2.5 N for glass plates). Through a PID control mechanism, the system compares the contact force fed back by the six-dimensional force sensor in real time with the target touch force, dynamically adjusting the driving torque of the gripping device motor to ensure the contact force remains stable within the allowable fluctuation range (e.g., ±0.15 N), preventing damage to the target placement position or sample plate due to excessive torque. By combining the physical property parameters of the material type (density, elastic modulus, surface friction coefficient, etc.), and based on the robot dynamics model and the mechanical structure parameters of the gripping device, a mathematical relationship between motor torque, touch force, and material properties is established. The second motor torque required to ensure stable placement is calculated. The second motor torque calculated by this model is sent to the drive motor of the gripping device to ensure that the sample plate is placed stably. When the force sensor detects a sudden change in contact force and the sample plate has been removed from the gripping device, the robotic arm is controlled to lift up.

[0081] During the lifting process of the robotic arm, the robotic arm's camera device captures an image of the target placement position. The placement result is determined through visual image analysis. If the image analysis results show that the sample plate has been accurately placed in the target position and the pose error is within the allowable range, the placement result is determined to be successful, and the robot control system feeds this result back to the host computer. If placement fails, an error code (such as E-002: placement offset) is automatically recorded, and a retry strategy is generated (up to 3 retries).

[0082] As can be seen, in this embodiment, through the coordinated efforts of multiple links such as visual positioning, empty space detection, dynamic parameter adjustment and contact force control, the precise control of the placement of the target sample plate is achieved, which improves the efficiency and adaptability of the robotic arm in laboratory automation and significantly enhances the reliability and efficiency of sample transfer in high-throughput experimental scenarios.

[0083] In one possible embodiment, the target sample plate is equipped with a sample label, which is used to display relevant information about the target sample plate. The empty space situation also includes a non-idle situation. The method further includes: obtaining label information of the target sample plate based on the sample label, the label information including relevant information of the target sample plate; determining an information comparison result based on the label information and the initialization task information, the comparison result including information consistency and information inconsistency; if the comparison result is that the information is consistent, then controlling the humanoid robot to return to the initial position; if the comparison result is that the information is inconsistent, then reporting an abnormal situation and controlling the humanoid robot to return to the initial position; if the first empty space situation is the non-idle situation, then reporting an abnormal situation and generating a retry strategy, controlling the humanoid robot to repeatedly obtain the second empty space situation of the target placement position based on the retry strategy; if the second empty space situation is the non-idle situation, controlling the humanoid robot to carry the target sample plate back to the initial position; if the second empty space situation is the idle situation, controlling the humanoid robot to perform a material placement operation; when it is determined that the placement result is that the position is occupied, controlling the humanoid robot to return to the initial position.

[0084] Before performing the material placement operation, if the first available slot is detected as not being idle, the system reports an exception (error code E-004) and generates a retry strategy. The retry strategy is based on a time window mechanism, for example, setting a retry every 30 seconds within a 3-minute period to avoid task failure due to temporary occupation. During the retry cycle, the robotic arm continuously acquires information on the second available slot: if it is still not idle, the robotic arm returns to its initial position with the sample plate, and the host computer marks the task as "pending manual processing" and notifies laboratory management personnel via push notification (e.g., SMS, email). If it becomes idle, the material placement operation is immediately performed to ensure efficient task completion.

[0085] The system detects that the first empty slot is vacant and that material placement has been completed. It then retrieves the label information of the target sample plate. Each sample plate can have a label for scanning, and the scanning result contains relevant information about that sample plate. This label information can be obtained through a head-mounted camera or a robotic arm camera. The retrieved label information is compared with the initial task information (such as the target plate ID and material type). If the sample ID in the label matches the target plate ID, and the material type is consistent with the task requirements, the system determines the comparison result as "information consistent" and triggers the robotic arm to return to its initial position. If a sample ID mismatch occurs (e.g., plate A is placed in plate B) or a material type conflict occurs (e.g., a liquid sample is mistakenly placed in a solid plate), the system immediately reports an anomaly (error code E-003) and alerts the operator via an audible and visual alarm (e.g., LED lights + buzzer). The anomaly details (e.g., error type, timestamp, location coordinates) are stored in the log system. After reporting the anomaly, the robotic arm returns to its initial position to prevent the error from escalating.

[0086] Once the system determines the placement result as "seat in place" through visual analysis, the robotic arm's camera device re-captures the target placement position. A template matching algorithm verifies the overlap between the sample plate edge and the plate position markings (with a required deviation ≤1mm). Simultaneously, force sensor data confirms that the contact force is stable within a preset range (e.g., 0.5N ± 0.1N). This dual verification ensures placement quality. Then, the humanoid robot is controlled to return to its initial position, and the system updates the task status to "completed."

[0087] As can be seen, in this embodiment, by comparing the sample label information with the initialization task information in real time, the matching between the sample plate and the target location can be accurately identified. When the target placement location is temporarily occupied (e.g., when researchers temporarily place items), the system continuously monitors the empty space status through a retry strategy to avoid task failure due to occasional conflicts. Through closed-loop control of "information verification - dynamic fault tolerance - quality verification - automatic regression", the accuracy, reliability, and efficiency of the sample plate transfer task are improved.

[0088] For example, in an antibody screening experiment at a biomedical laboratory, centrifuged sample plates (including 96-well and 384-well plates) need to be transferred from a high-speed centrifuge to an ELISA reader for fluorescence detection. The experimental procedure includes the following key steps: the centrifuge has 8 plate positions, a maximum speed of 15,000 rpm, and the sample plates must be detected within 30 minutes after centrifugation. The ELISA reader has 12 plate positions, supports multi-wavelength detection, and must be strictly matched to the sample plate type. The laboratory environment constraints include a laboratory area of ​​60㎡, containing 12 pieces of equipment such as pipettes and incubators, with a passageway width of only 1.2m. The task initialization phase involves task reception and parsing. Researchers configure the task list via a graphical interface and send instructions to the controller, including: the first target location is the centrifuge (ID: LXJ-01), and the second target location is the ELISA reader (ID: MTP-02); the target sample plate information includes three 96-well plates (Type A) and two 384-well plates (Type B), with Type A plates given high priority due to the presence of unstable antibody markers; the time constraint is to complete the testing within 30 minutes of centrifugation. After the humanoid robot starts, it performs a 360° scan of the laboratory using its head-mounted 3D camera, generating a point cloud map containing equipment locations and aisle widths in 45 seconds. Key areas are marked: centrifuge door location (X=2.5m, Y=1.8m), ELISA reader plate area (X=4.2m, Y=3.5m), and high-temperature equipment area (X=1.2m, Y=0.8m to X=1.8m, Y=1.5m). The A* algorithm is used to generate a path from the initial position (X=0, Y=0) to the first target position, i.e., the centrifuge. The robot travels straight along the main aisle, then turns 90° to the right, decelerates to 0.5 m / s to pass through a narrow area, and finally reaches a point 1 m in front of the centrifuge. The pre-planned robotic arm trajectory descends from a safe height (Z=1.2 m) to the plate gripping height (Z=0.3 m), with the end effector adjusted to -45° (to accommodate the centrifuge's tilting door). The humanoid robot travels along the planned path at a speed of 1.2 m / s, with a LiDAR scanner scanning for obstacles in real time. When a pipette (moving speed 0.6 m / s) enters the safe distance (1.5 m), the DWA algorithm is triggered: predicting a collision time of 0.8 seconds, adjusting the speed to 0.8 m / s, and simultaneously shifting 0.2 m to the right to maintain a 1 m safe distance. Bipedal stabilization: Before reaching the centrifuge, the robot switches to bipedal mode, and the IMU detects the ground flatness (tilt <3°). The robotic arm is pre-raised to Z=1.0m to avoid collision with the hatch. The robotic arm's camera focuses on the target plate position (e.g., the 3rd position on the centrifuge), and identifies the first positioning code, i.e., the plate position QR code, using the YOLOv8 model to obtain the precise coordinates (X=2.55m, Y=1.82m, Z=0.28m). Combined with the plate position offset compensation algorithm after centrifugation (based on historical data, with an average offset of 0.3mm), the target position of the gripping device is corrected.The gripping device descends at a speed of 0.3 m / s, triggering force control mode upon contact with the sample plate. The initial force threshold is 0.5 N (contact detection), followed by progressive gripping from 1.5 N (plastic plate) to 2.0 N (stabilizing after 3 seconds). A six-dimensional force sensor monitors torque changes in real time, adjusting the robotic arm's posture to balance the sample plate's center of gravity (maximum offset 5 mm). After material gripping, the humanoid robot moves towards the microplate reader (second target position) at a speed of 0.8 m / s. An obstacle is detected en route, and the incubator door opens (occupying 1 / 3 of the passage). An obstacle avoidance path is generated using the RRT* algorithm, taking 1.2 seconds. The speed decreases to 0.5 m / s during obstacle avoidance to ensure the sample plate's horizontal deviation is <0.8°. The robotic arm uses a zero-force control mode, compensating for chassis vibration at 5-15 Hz via a joint spatial impedance controller, with a maximum amplitude compensation of ±2 mm. Upon reaching the microplate reader, the robotic arm's camera scans the second positioning code (the plate placement QR code) again, revealing a cumulative positioning error of 2.8 mm (mainly caused by uneven ground). Visual servo control is triggered, and the robotic arm's end effector fine-tunes to the target position (X = 4.20 ± 0.005 m, Y = 3.50 ± 0.005 m). The gripper descends at 0.2 m / s, applying 0.8 N of pressure for 1 second upon contact with the microplate reader. The force sensor detects this pressure change and triggers a release command, causing the gripper to withdraw at 0.1 m / s. After the target sample plate is placed, the reader reads the sample label (ID: AB-0012) and compares it with the task list on the host computer, initializing the task information: plate type (96 wells), experimental item (antibody screening), and expiration date (2024-12-31). Then, the plate placement status is confirmed: the microplate reader's photoelectric sensor detects the sample plate is in place (signal level changes from high to low). Then, the humanoid robot sends a task completion signal (including a timestamp: 14:23:17) to the host computer.

[0089] In some cases, after the target sample plate is placed, the RFID reader at the second target location can also read the sample plate label (ID: AB-0012) using its built-in RFID reader and compare it with the task list on the host computer. That is, each plate position can be equipped with a tool for scanning sample labels, and each plate position communicates with the system to achieve information sharing.

[0090] It should be noted that the tasks sent from the host computer to the controller may include multiple subtasks, that is, multiple task information. Priority needs to be determined, and each task is completed in descending order of priority.

[0091] In one possible embodiment, please refer to Figure 3 , Figure 3 This is a schematic flowchart of another intelligent laboratory sample plate transfer control method provided in this application embodiment, as shown below. Figure 3As shown, the process begins with task initialization and movement. The controller receives task instructions from the host computer, parses them to obtain initial task information, including the first target position and the second target position, etc. It then uses SLAM to build an environmental map and plan a path, controlling the robot to move from the initial position to the first target position. During movement, obstacle avoidance or path replanning is performed. After reaching the first target position, precise grasping preparation is initiated. The end-effector camera identifies the first positioning code of the source board station, obtaining sub-millimeter-level coordinate information. A six-dimensional force sensor and IMU detect joint vibration and damping parameters, calculate grasping error compensation values ​​(such as the target damping coefficient), and correct the end-effector trajectory. Adaptive grasping and fine-tuning then occur. The gripper descends at a speed of 0.3 m / s, triggering force control mode upon contact with the sample board. The gripping force (1.5 N / 2.5 N) is adjusted according to the material type (e.g., plastic / glass). The IMU monitors center-of-gravity shift and dynamically adjusts the arm posture to ensure stable grasping. After successful acquisition, the humanoid robot is moved to the second target location. During the transfer, dynamic control, obstacle avoidance, and path replanning are performed. The robot switches to wheeled mode for movement, and the LiDAR detects obstacles in real time, triggering the RRT* algorithm to replan the path. The PID algorithm compensates for chassis vibration, and the vision system monitors the sample plate's levelness (fine-tuning when the deviation is >0.5°). Upon reaching the second target location, verification and precise placement are performed. The camera identifies the second positioning code, and the photoelectric sensor detects the plate's vacancy status. If vacant, precise placement is performed, dynamically adjusting damping and torque based on mechanical data to avoid impact. Visual confirmation confirms the sample plate is fully seated (deviation ≤1mm). The sample label information is read and compared with the task information; if they do not match, an anomaly is reported. If the target plate is not vacant, the system automatically retryes n detections until vacant or returns to the initial position, where n is a preset number of attempts. If vacant is found within n attempts, precise placement is performed, dynamically adjusting damping and torque based on mechanical data to avoid impact. Visual confirmation confirms the sample plate is fully seated (deviation ≤1mm). The system reads the sample label information and compares it with the task information. If they do not match, an anomaly is reported. If there is no free time within n attempts, an anomaly is reported and the system returns to its original position. After the task is completed, the system controls the humanoid robot to return to its initial position.

[0092] Please see Figure 4 , Figure 4 This is a schematic diagram of a sample plate transfer control device for an intelligent laboratory proposed in this application embodiment. The device is applied to the controller of a humanoid robot system. The humanoid robot system further includes the humanoid robot itself, which includes a robotic arm equipped with a gripping device for grasping materials, including target sample plates. The humanoid robot system also includes a plate station module and an environmental perception and navigation system. The plate station module includes a first target position and a second target position. Figure 4 As shown, the sample plate transfer control device 400 of the intelligent laboratory includes:

[0093] An initialization module 410 is used to acquire task information and determine initialization task information based on the task information, and to control the humanoid robot to move from an initial position to the first target position based on the initialization task information.

[0094] The first material grasping module 420 is used to acquire first sensing data of the humanoid robot at the first target position from the environmental perception and navigation system, and determine a grasping error compensation value based on the first sensing data. The first sensing data includes first visual data and first mechanical data.

[0095] The second material gripping module 430 is used to control the humanoid robot to perform a material gripping operation on the target sample plate based on the gripping error compensation value, and to obtain a first feedback signal. Based on the first feedback signal, the module makes fine adjustments to enable the humanoid robot to complete the material gripping operation. The first feedback signal is second visual data and / or second mechanical data detected during the material gripping operation.

[0096] The material transfer module 440 is used to detect the material grasping status. If the material grasping status is detected as a completed grasping status, the humanoid robot is controlled to move to the second target position based on the initialization task information. During the movement, the status is dynamically adjusted based on the second feedback information. The second feedback information includes the third visual data and / or the third mechanical data detected during the movement.

[0097] The material placement module 450 is used to obtain third feedback information and place the target sample plate based on the third feedback information when the humanoid robot is detected to have reached the second target position. The third feedback information includes fourth visual data and / or fourth mechanical data.

[0098] In one possible embodiment, the first visual data comes from the robotic arm's camera device, and the first material grasping module 420, in acquiring the first sensing data of the humanoid robot at the first target position and determining the grasping error compensation value based on the first sensing data, is specifically used for:

[0099] Based on the first visual data, obtain the coordinate information of the first board station based on the first positioning code. The first positioning code is the positioning code of the first target location. The coordinate information of the first board station is used to indicate the precise coordinates of the board station where the first target location is located.

[0100] Based on the first mechanical data, a first damping parameter and a first vibration frequency parameter are determined. The first damping parameter and the first vibration frequency parameter are used to characterize the real-time vibration and damping of the joints of the humanoid robot.

[0101] A first error compensation value is determined based on the first damping parameter and the first plate station coordinate information. The first error compensation value includes a first target damping coefficient, which is used to adjust and suppress the error caused by the vibration of the robotic arm.

[0102] In one possible embodiment, the second visual data in the first feedback signal comes from the robotic arm camera device, and the initialization task information further includes a material type, which is used to indicate the characteristics of the target sample plate; the second material gripping module 430, in acquiring the first feedback signal and performing fine-tuning based on the first feedback signal to complete the material gripping operation, is specifically used for:

[0103] Continuously acquire the first feedback signal during the crawling process;

[0104] If the first feedback signal is detected to meet the first contact condition, the touch intensity is adjusted based on the material type. The first contact condition is used to constrain the relative distance and / or contact force between the humanoid robot and the target sample plate.

[0105] The torque of the first motor is determined based on the material type and the touch force, and the material gripping operation is completed based on the torque of the first motor.

[0106] The current center of gravity parameter is determined based on the first feedback signal. The center of gravity parameter is used to characterize the center of gravity of the target sample plate relative to the gripper of the humanoid robot.

[0107] The torque of the second motor is determined based on the center of gravity parameters, and the center of gravity offset is continuously detected. The center of gravity offset is used to characterize the difference between the center of gravity parameters and the target center of gravity parameters.

[0108] The center of gravity is adjusted based on the aforementioned offset to smoothly complete the material gripping operation.

[0109] In one possible embodiment, the third visual data comes from the head camera device and / or the robotic arm camera device, and the material transfer module 440 is specifically used for: dynamically adjusting the state based on the second feedback information during the movement.

[0110] Receive second feedback information from the environmental perception and navigation system;

[0111] When a sudden obstacle is detected based on the second feedback information, the humanoid robot is controlled to brake urgently, and the path is replanned based on the environmental perception and navigation system; and, chassis vibration compensation data is determined based on the PID algorithm, and levelness data is determined based on the second feedback information. The chassis vibration compensation data is used to reduce the degree of chassis vibration of the humanoid robot, and the levelness data is the levelness data of the target sample plate.

[0112] The humanoid robot's state is dynamically adjusted based on the chassis vibration compensation data and the levelness data.

[0113] In one possible embodiment, the fourth visual data comes from the robotic arm camera device, and the material placement module 450 is specifically used for: acquiring the third feedback information and placing the target sample plate based on the third feedback information.

[0114] Based on the fourth visual data, the coordinate information of the second board station based on the second positioning code is obtained. The second positioning code is the positioning code of the second target position. The coordinate information of the second board station is used to indicate the precise coordinates of the board station where the first target position is located.

[0115] Obtain the first empty space status of the target placement location, wherein the target placement location is the empty space indicated by the initialization task information for placing the target sample plate, and the first empty space status includes the idle status;

[0116] If the first empty space is the idle state, the humanoid robot is controlled to perform a material placement operation. During the material placement operation, a second damping parameter and a second vibration frequency parameter are determined based on the fourth mechanical data. The second damping parameter and the second vibration frequency parameter are used to characterize the real-time vibration and damping of the humanoid robot's joints. A second error compensation value is determined based on the second damping parameter and the second plate coordinate information. The second error compensation value includes a second target damping coefficient, which is used to adjust and suppress the error caused by the vibration of the robotic arm. A third feedback signal is continuously acquired during the placement process.

[0117] If the third feedback signal is detected to meet the second contact condition, the touch intensity is adjusted based on the material type. The second contact condition is used to constrain the relative distance and / or contact force between the humanoid robot and the target placement position.

[0118] The torque of the second motor is determined based on the material type and the touch force, and the material placement operation is completed based on the torque of the second motor.

[0119] When it is detected that the target sample plate has been detached from the gripping device, the robotic arm is controlled to lift up and acquire the target placement position image of the robotic arm camera device;

[0120] The placement result is determined based on the target location placement image, and the placement result includes being seated.

[0121] In one possible embodiment, the target sample plate is equipped with a sample label, which is used to display relevant information about the target sample plate. The empty space condition also includes a non-empty space condition. The material placement module 450 is further used for:

[0122] The label information of the target sample plate is obtained based on the sample label, and the label information includes relevant information of the target sample plate;

[0123] Based on the tag information and the initialization task information, an information comparison result is determined, including information matching and information inconsistency; if the comparison result is that the information is matching, the humanoid robot is controlled to return to the initial position; if the comparison result is that the information is inconsistent, an abnormal situation is reported and the humanoid robot is controlled to return to the initial position.

[0124] If the first empty space condition is a non-empty condition, an abnormal situation is reported and a retry strategy is generated. Based on the retry strategy, the humanoid robot is controlled to repeatedly obtain the second empty space condition of the target placement position. If the second empty space condition is a non-empty condition, the humanoid robot is controlled to carry the target sample plate back to the initial position. If the second empty space condition is an empty condition, the humanoid robot is controlled to perform material placement operation.

[0125] Once the placement result is confirmed as "seat is in place", control the humanoid robot to return to the initial position.

[0126] It is worth noting that the specific functional implementation of the sample plate transfer control device in the intelligent laboratory is described above. Figure 2 The description of the sample plate transfer control method in the intelligent laboratory illustrates, for example, the initialization module used to implement the relevant content of S210. The various units or modules in the sample plate transfer control device 400 of the intelligent laboratory can be individually or entirely merged into one or more other units or modules, or some of the units or modules can be further divided into multiple functionally smaller units or modules. This achieves the same operation without affecting the technical effects of the embodiments of the present invention. The aforementioned units or modules are based on logical function division. In practical applications, the function of one unit (or module) is implemented by multiple units (or modules), or the function of multiple units (or modules) is implemented by one unit (or module).

[0127] In the case of using integrated units, please refer to Figure 5 , Figure 5 This is a schematic diagram of another intelligent laboratory sample plate transfer control device provided in this application embodiment, as shown below. Figure 5 As shown, the intelligent laboratory sample plate transfer control device 400 includes a processing module 402 and a communication module 401. The processing module 402 controls and manages the actions of the intelligent laboratory sample plate transfer control device 400, for example, executing the steps of the initialization module 410, the first material gripping module 420, the second material gripping module 430, the material transfer module 440, and the material placement module 450, and / or executing other processes of the technology described herein. The communication module 401 is used for interaction between the intelligent laboratory sample plate transfer control device 400 and other devices. Figure 5 As shown, the intelligent laboratory sample plate transfer control device 400 may also include a storage module 403, which is used to store the program code and data of the intelligent laboratory sample plate transfer control device 400.

[0128] The processing module 402 can be a processor or controller, such as a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc. The communication module 401 can be a transceiver, RF circuitry, or a communication interface, etc. The storage module 403 can be a memory.

[0129] All relevant content in each scenario involved in the above method embodiments can be referenced from the functional descriptions of the corresponding functional modules, and will not be repeated here. The sample plate transfer control device 400 of the above-mentioned intelligent laboratory can perform the above-mentioned... Figure 2 The sample plate transfer control method of the intelligent laboratory shown is illustrated.

[0130] As can be seen, the intelligent laboratory sample plate transfer control device described in this application embodiment acquires task information and determines initialization task information based on the task information, and controls the humanoid robot to move from an initial position to the first target position based on the initialization task information; acquires first sensing data of the humanoid robot at the first target position from the environmental perception and navigation system, and determines a grasping error compensation value based on the first sensing data, the first sensing data including first visual data and first mechanical data; controls the humanoid robot to perform material grasping operation on the target sample plate based on the grasping error compensation value, and acquires a first feedback signal, and makes fine adjustments based on the first feedback signal to make the humanoid robot... The material grasping operation is completed, and the first feedback signal is the second visual data and / or second mechanical data detected during the material grasping operation. The material grasping state is detected; if the material grasping state is detected as completed, the humanoid robot is controlled to move to the second target position based on the initialization task information, and its state is dynamically adjusted based on the second feedback information during the movement. The second feedback information includes the third visual data and / or third mechanical data detected during the movement. When the humanoid robot is detected to have reached the second target position, the third feedback information is acquired and the target sample plate is placed based on the third feedback information, which includes the fourth visual data and / or fourth mechanical data. Thus, through scenario-based compliant control and vision-force coordinated control, the efficiency and adaptability of the robotic arm in laboratory automation are improved, significantly enhancing the reliability and efficiency of sample transfer in high-throughput experimental scenarios.

[0131] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of an electronic device proposed in an embodiment of this application. As shown in the figure, the electronic device 600 includes a processor 610, a memory 620, a communication interface 630, and one or more programs 621. The one or more programs 621 are stored in the memory 620 and are configured to be executed by the processor 610.

[0132] The processor 610, memory 620, and communication interface 630 are interconnected and perform communication between them.

[0133] The memory 620 can be volatile memory such as dynamic random access memory (DRAM) or non-volatile memory such as a hard disk drive (HDD). The memory 620 stores a set of executable program code, and the processor 610 calls one or more programs 621 stored in the memory 620 to execute the above-described program. Figure 2 Some or all of the steps of the sample plate transfer control method for any intelligent laboratory described in the embodiments.

[0134] Among them, electronic device 600 may include controllers, etc. The above is only an example and not an exhaustive list, including but not limited to the above electronic devices.

[0135] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.

[0136] This application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program operable to cause a computer to perform some or all of the steps of any of the methods described in the above method embodiments. The computer program product may be a software installation package, and the computer may include an electronic device.

[0137] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0138] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0139] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0140] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0141] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0142] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer electronic device (which may be a personal computer, electronic device, or network electronic device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0143] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0144] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A sample plate transfer control method for an intelligent laboratory, characterized in that, A controller for a humanoid robot system, the humanoid robot system further comprising the humanoid robot, the humanoid robot including a robotic arm equipped with a gripping device for grasping materials, the materials including a target sample plate, the humanoid robot system further comprising a plate station module and an environmental perception and navigation system, the plate station module including a first target position and a second target position; the method comprising: Acquire task information and determine initialization task information based on the task information, and control the humanoid robot to move from the initial position to the first target position based on the initialization task information; The first sensing data of the humanoid robot at the first target position is obtained from the environmental perception and navigation system, and the grasping error compensation value is determined based on the first sensing data. The first sensing data includes first visual data and first mechanical data. Based on the grasping error compensation value, the humanoid robot is controlled to perform a material grasping operation on the target sample plate, and a first feedback signal is obtained. Based on the first feedback signal, the humanoid robot is fine-tuned to complete the material grasping operation. The first feedback signal is the second visual data and / or the second mechanical data detected during the material grasping operation. The material grasping status is detected. If the material grasping status is detected as completed, the humanoid robot is controlled to move to the second target position based on the initialization task information. During the movement, the status is dynamically adjusted based on the second feedback information. The second feedback information includes the third visual data and / or the third mechanical data detected during the movement. When the humanoid robot is detected to have reached the second target position, third feedback information is obtained and the target sample plate is placed based on the third feedback information. The third feedback information includes fourth visual data and / or fourth mechanical data.

2. The method according to claim 1, characterized in that, The first visual data comes from the robotic arm's camera device. Acquiring the first sensor data of the humanoid robot at the first target position and determining the grasping error compensation value based on the first sensor data includes: Based on the first visual data, obtain the coordinate information of the first board station based on the first positioning code. The first positioning code is the positioning code of the first target location. The coordinate information of the first board station is used to indicate the precise coordinates of the board station where the first target location is located. Based on the first mechanical data, a first damping parameter and a first vibration frequency parameter are determined. The first damping parameter and the first vibration frequency parameter are used to characterize the real-time vibration and damping of the joints of the humanoid robot. A first error compensation value is determined based on the first damping parameter and the first plate station coordinate information. The first error compensation value includes a first target damping coefficient, which is used to adjust and suppress the error caused by the vibration of the robotic arm.

3. The method according to claim 1 or 2, characterized in that, The second visual data in the first feedback signal comes from the robotic arm camera device. The initialization task information also includes the material type, which is used to indicate the characteristics of the target sample plate. Acquiring the first feedback signal and fine-tuning based on the first feedback signal to complete the material grasping operation includes: Continuously acquire the first feedback signal during the crawling process; If the first feedback signal is detected to meet the first contact condition, the touch intensity is adjusted based on the material type. The first contact condition is used to constrain the relative distance and / or contact force between the humanoid robot and the target sample plate. The torque of the first motor is determined based on the material type and the touch force, and the material gripping operation is completed based on the torque of the first motor. The current center of gravity parameter is determined based on the first feedback signal. The center of gravity parameter is used to characterize the center of gravity of the target sample plate relative to the gripper of the humanoid robot. The torque of the second motor is determined based on the center of gravity parameters, and the center of gravity offset is continuously detected. The center of gravity offset is used to characterize the difference between the center of gravity parameters and the target center of gravity parameters. The center of gravity is adjusted based on the aforementioned offset to smoothly complete the material gripping operation.

4. The method according to claim 1, characterized in that, The robotic arm is equipped with a robotic arm camera device, and the humanoid robot also includes a head, which is equipped with a head camera device. The third visual data comes from the head camera device and / or the robotic arm camera device. The dynamic state adjustment based on second feedback information during movement includes: Receive second feedback information from the environmental perception and navigation system; When a sudden obstacle is detected based on the second feedback information, the humanoid robot is controlled to brake urgently, and the path is replanned based on the environmental perception and navigation system; and, chassis vibration compensation data is determined based on the PID algorithm, and levelness data is determined based on the second feedback information. The chassis vibration compensation data is used to reduce the degree of chassis vibration of the humanoid robot, and the levelness data is the levelness data of the target sample plate. The humanoid robot's state is dynamically adjusted based on the chassis vibration compensation data and the levelness data.

5. The method according to claim 1, characterized in that, The fourth visual data comes from the robotic arm camera device, and the step of acquiring the third feedback information and placing the target sample plate based on the third feedback information includes: Based on the fourth visual data, the coordinate information of the second board station based on the second positioning code is obtained. The second positioning code is the positioning code of the second target location. The coordinate information of the second board station is used to indicate the precise coordinates of the board station where the second target location is located. Obtain the first empty space status of the target placement location, wherein the target placement location is the empty space indicated by the initialization task information for placing the target sample plate, and the first empty space status includes the idle status; If the first empty space is the idle state, the humanoid robot is controlled to perform a material placement operation. During the material placement operation, a second damping parameter and a second vibration frequency parameter are determined based on the fourth mechanical data. The second damping parameter and the second vibration frequency parameter are used to characterize the real-time vibration and damping of the humanoid robot's joints. A second error compensation value is determined based on the second damping parameter and the second plate coordinate information. The second error compensation value includes a second target damping coefficient, which is used to adjust and suppress the error caused by the vibration of the robotic arm. A third feedback signal is continuously acquired during the placement process. If the third feedback signal is detected to meet the second contact condition, the touch intensity is adjusted based on the material type. The second contact condition is used to constrain the relative distance and / or contact force between the humanoid robot and the target placement position. The torque of the second motor is determined based on the material type and the touch force, and the material placement operation is completed based on the torque of the second motor. When it is detected that the target sample plate has been detached from the gripping device, the robotic arm is controlled to lift up and acquire the target placement position image of the robotic arm camera device; The placement result is determined based on the target location placement image, and the placement result includes being seated.

6. The method according to claim 5, characterized in that, The target sample plate is equipped with a sample label, which is used to display relevant information about the target sample plate. The empty space also includes a non-empty space. The method further includes: The label information of the target sample plate is obtained based on the sample label, and the label information includes relevant information of the target sample plate; Based on the tag information and the initialization task information, an information comparison result is determined, including information matching and information inconsistency; if the comparison result is that the information is matching, the humanoid robot is controlled to return to the initial position; if the comparison result is that the information is inconsistent, an abnormal situation is reported and the humanoid robot is controlled to return to the initial position. If the first empty space condition is a non-empty condition, an abnormal situation is reported and a retry strategy is generated. Based on the retry strategy, the humanoid robot is controlled to repeatedly obtain the second empty space condition of the target placement position. If the second empty space condition is a non-empty condition, the humanoid robot is controlled to carry the target sample plate back to the initial position. If the second empty space condition is an empty condition, the humanoid robot is controlled to perform material placement operation. Once the placement result is confirmed as "seat is in place", control the humanoid robot to return to the initial position.

7. A humanoid robot system applying the sample plate transfer control method of an intelligent laboratory as described in any one of claims 1-6, characterized in that, include: A humanoid robot, comprising a mobile chassis and a robotic arm, wherein the robotic arm is equipped with a gripping device for gripping materials, the materials including the target sample plate, and the mobile chassis comprising a bipedal structure and / or a wheeled structure, wherein the humanoid robot switches to the wheeled structure when moving and switches to the bipedal structure when standing. A controller for performing the method as described in any one of claims 1-6; An environmental perception and navigation system, comprising a vision system, a six-dimensional force sensor, and a navigation and positioning system, wherein the vision system is used to acquire visual data, the six-dimensional force sensor is used to acquire mechanical data, and the navigation and positioning system is used to provide path information and the intelligent laboratory space information; The plate station module is a multi-modal area for storing sample plates, and the plate station module is a multi-modal design.

8. The system according to claim 7, characterized in that, The vision system includes multiple camera devices, including a head camera device and a robotic arm camera device; The six-dimensional force sensor includes a robotic arm force sensor for sensing the force on the gripping device and / or the end of the robotic arm; the board station module is a multi-layer board station built with a C-shaped channel steel frame, and the board station module includes an operating area, with at least one positioning code set in each operating area; The controller, the head camera device, the robotic arm camera device, and the six-dimensional force sensor are synchronized in time and space; The humanoid robot's robotic arm includes a robotic arm with 7 degrees of freedom.

9. A sample plate transfer control device for an intelligent laboratory, characterized in that, A controller for a humanoid robot system, the humanoid robot system further including the humanoid robot, the humanoid robot including a robotic arm, the robotic arm being equipped with a gripping device for gripping materials, the materials including a target sample plate, the humanoid robot system further including a plate station module and an environmental perception and navigation system, the plate station module including a first target position and a second target position; the device includes: An initialization module is used to acquire task information and determine initialization task information based on the task information, and to control the humanoid robot to move from an initial position to the first target position based on the initialization task information; The first material grasping module is used to acquire first sensing data of the humanoid robot at the first target position from the environmental perception and navigation system, and determine the grasping error compensation value based on the first sensing data. The first sensing data includes first visual data and first mechanical data. The second material grasping module is used to control the humanoid robot to perform material grasping operation on the target sample plate based on the grasping error compensation value, and to obtain a first feedback signal. Based on the first feedback signal, the module makes fine adjustments to enable the humanoid robot to complete the material grasping operation. The first feedback signal is second visual data and / or second mechanical data detected during the material grasping operation. The material transfer module is used to detect the material grasping status. If the material grasping status is detected as completed, the module controls the humanoid robot to move to the second target position based on the initialization task information. During the movement, the module dynamically adjusts the status based on the second feedback information. The second feedback information includes the third visual data and / or the third mechanical data detected during the movement. The material placement module is used to obtain third feedback information and place the target sample plate based on the third feedback information when the humanoid robot is detected to have reached the second target position. The third feedback information includes fourth visual data and / or fourth mechanical data.

10. An electronic device, characterized in that, The method includes a processor and a memory storing execution instructions, the memory storing one or more programs; when the processor executes the execution instructions stored in the memory, the processor performs the method as described in any one of claims 1 to 6.

Images