Robotic control method and system for medical laboratory standardized work

Abstract

The application relates to the technical field of medical laboratory information processing, and discloses a somatized robot control method and system for standardized operation of a medical laboratory. The method comprises the following steps: acquiring multi-modal state data of a target operation object, operation equipment, an operation space and a somatized robot in the medical laboratory, and updating a corresponding digital twin based on the multi-modal state data; determining a current compliance state of the target operation object and a candidate action set allowed to be executed by the somatized robot according to the digital twin and operation rules of the medical laboratory; constructing an executable task graph according to a business process corresponding to the target operation object, task constraints, running states of the operation equipment and the somatized robot and the candidate action set, generating a target task sequence; controlling the somatized robot to execute corresponding standardized operation, updating the digital twin and the compliance state according to an execution result, and forming a trace evidence chain, so that the standardized operation of a specimen, reagent consumables and dangerous goods is closed-loop controlled.

Description

Technical Field

[0001] This application relates to the field of medical laboratory information processing technology, and in particular to an embodied robot control method and system for standardized operations in medical laboratories. Background Technology

[0002] Medical laboratories undertake a large volume of specimen testing, verification, preservation, reagent and consumable management, and hazardous materials handling. Many standardized, repetitive tasks in medical laboratories, such as specimen receipt, barcode scanning, centrifugation balancing, decapping, loading, recapping, transport, preservation, waste disposal, reagent receipt, warehousing, outbound, opening, labeling, inventory, and hazardous materials transport, still rely primarily on manual labor. While these tasks are not technically demanding, they are frequent, time-consuming, and carry high risks of error and biosafety, requiring complete records that meet quality management and review requirements.

[0003] Although the existing market offers automated pretreatment systems, posttreatment systems, fully automated laboratory production lines, single capping robots, sample sorting and storage systems, and reagent inventory management systems, these systems are insufficient to cover the entire process of medical laboratory operations, including reagents and consumables, hazardous waste, cold chain, contact records, and review evidence. Furthermore, they are difficult to integrate with LIS (Laboratory Information System), LIMS (Laboratory Information Management System), RIMS (Reagent Information Management System), quality systems, and clinical review rules, which can easily lead to risks of mis-taking, misplacement, omissions, or oversights in processing or recording. Summary of the Invention

[0004] Therefore, it is necessary to provide a holographic robot control method and system for standardized operations in medical laboratories to address the aforementioned technical problems.

[0005] In a first aspect, this application provides a method for controlling an embodied robot for standardized operations in medical laboratories, the method comprising: Acquire multimodal state data of target work objects, work equipment, work space, and embodied robots in a medical laboratory, and update the digital twins corresponding to the target work objects, work equipment, work space, and embodied robots based on the multimodal state data; Based on the digital twin and preset medical laboratory operating rules, the current compliance status of the target operating object is determined, and based on the current compliance status and the action authorization conditions associated with the current compliance status, a set of candidate actions that the embodied robot is allowed to perform is determined; Based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, and the candidate action set, an executable task graph is constructed, and a target task sequence is generated according to the executable task graph. The robot is controlled to perform standardized tasks for the target object at different workstations according to the target task sequence, and the execution results are obtained. The digital twin and the current compliance status are updated based on the execution results, and the multimodal status data, compliance status changes, action parameters and execution results are associated to form a traceability evidence chain for the standardized operation.

[0006] In one example, the method further includes a process for creating the digital twin, the process of creating the digital twin including: Data association and state parsing are performed on the multimodal state data to obtain the object state of the target operation object, the operating state of the operation equipment and the embodied robot, and the environmental state of the operation space; The object status, running status, and environment status are written into the status field of the corresponding digital twin, and the current job status of the target job object is associated with the corresponding job process node according to the business process corresponding to the target job object to obtain the digital twin.

[0007] In one example, based on the digital twin and preset medical laboratory operating rules, the current compliance status of the target operating object is determined, and based on the current compliance status and the action authorization conditions associated with the current compliance status, a set of candidate actions that the embodied robot is allowed to perform is determined, including: The medical laboratory operation rules are converted into status recognition conditions and action authorization conditions; The object status of the target operation object in the digital twin is matched with the status identification conditions to determine the current compliance status of the target operation object; Based on the current compliance status, query the associated action authorization conditions to obtain the allowed and prohibited actions under the current compliance status; The prohibited actions are excluded from the preset set of robot actions, and the remaining allowed actions are determined as the set of candidate actions.

[0008] In one example, the step of constructing an executable task graph based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, and the candidate action set, and generating a target task sequence based on the executable task graph, includes: The business process corresponding to the target job object is decomposed into multiple task nodes, and task attributes are configured for each task node; wherein, the task attributes include target job object, previous state, target state, target workstation, candidate action, completion time limit and exception handling conditions; Based on the execution sequence and compliance status transition relationships between each task node, task association edges are established to obtain the initial task graph; Based on the operating status of the work equipment and the embodied robot, the set of candidate actions, and the task constraints, the executable status of each task node in the initial task graph is determined, and the executable task graph is formed. The target task sequence is extracted from the executable task graph based on the preset task priority and the execution order represented by the task association edges.

[0009] In one example, determining the executable state of each task node in the initial task graph based on the operating status of the working equipment and the android, the candidate action set, and the task constraints, to form the executable task graph, includes: For each task node in the initial task graph, determine whether the target action corresponding to the task node belongs to the candidate action set, and determine whether the operating equipment and embodied robot required to perform the target action are available based on the operating status of the operating equipment and embodied robot. When the target action belongs to the candidate action set, the required work equipment and embodied robot are available, and the task node satisfies the corresponding task constraints, the task node is determined to be in an executable state; otherwise, the task node is determined to be in a waiting state or a blocked state. The task nodes and task-related edges in the initial task graph are updated according to the executable status of each task node to form the executable task graph.

[0010] In one example, the control of the embodied robot to perform standardized tasks for the target task object at different workstations according to the target task sequence, and to obtain the execution results, including: Based on the target work object, target action and target workstation in the target task sequence, determine the target execution device of the embodied robot, and determine the target end effector from multiple replaceable end effectors; Based on the structure and current spatial location of the target work object, generate the initial motion parameters of the target execution device and the target end effector; The corresponding target task is executed according to the initial action parameters, and execution feedback results are obtained during the execution process; wherein, the execution feedback results include at least one of visual feedback, position feedback, force feedback and collision feedback; Adjust the initial action parameters based on the execution feedback results, and after completing the standardized operation of the current target task, jump to the next task in the target task sequence until all operations in the target task sequence are completed, and obtain the state of the target task object after execution; The execution result is determined by comparing the status of the target job object after execution with the expected completion status of the target task. The result is determined as successful, unsuccessful, or abnormal.

[0011] In one example, updating the digital twin and the current compliance status based on the execution result, and associating the multimodal status data, compliance status changes, action parameters, and execution results to form a traceability evidence chain for the standardized operation, includes: The actual task completion status corresponding to the target task object is determined based on the execution result. Write the actual task completion status into the digital twin; Match the actual task completion status with the status transition conditions corresponding to the current compliance status; When the state transition condition is met, the current compliance state is transitioned to the corresponding next compliance state; When the state transition conditions are not met or the execution result is execution failure or execution exception, the target job object is converted to an abnormal suspension state or a state awaiting manual review, and an exception handling task is generated. Based on the updated content, compliance status transition information, and action parameters of the digital twin, an evidence chain record corresponding to the target operation object is generated, and the association between multiple evidence chain records is established according to the execution order of each target task in the target task sequence to form a traceability evidence chain for the standardized operation.

[0012] Secondly, the present invention also provides an embodied robot control system for standardized operations in medical laboratories, the system comprising: The twin update module is used to acquire multimodal state data of the target work object, work equipment, work space and embodied robot in the medical laboratory, and update the digital twin corresponding to the target work object, work equipment, work space and embodied robot based on the multimodal state data; The candidate action determination module is used to determine the current compliance status of the target operation object based on the digital twin and preset medical laboratory operation rules, and to determine a set of candidate actions that the embodied robot is allowed to perform based on the current compliance status and the action authorization conditions associated with the current compliance status. The task sequence generation module is used to construct an executable task graph based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, and the candidate action set, and to generate a target task sequence based on the executable task graph. The control execution module is used to control the embodied robot to perform standardized tasks for the target task object in different workstations according to the target task sequence, and to obtain the execution results; The evidence chain generation module is used to update the digital twin and the current compliance status based on the execution result, and associate the multimodal status data, compliance status changes, action parameters and execution results to form a traceability evidence chain for the standardized operation.

[0013] Thirdly, this application provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method provided in the first aspect of this application.

[0014] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method provided in the first aspect of this application.

[0015] The aforementioned embodied robot control method and system based on standardized operations in medical laboratories maps the target work object, work equipment, work space, and the multimodal state of the embodied robot to a digital twin. This enables the system to grasp the actual operational stage of specimens, reagents, consumables, or hazardous materials, as well as the status of surrounding equipment and the robot. Furthermore, it utilizes medical laboratory operation rules to determine the current compliance status and implements authorization gating for robot actions, ensuring that the robot only obtains the corresponding candidate action when the object status, quality rules, and operating permissions all meet the requirements. On this basis, it constructs an executable task graph and generates a target task sequence based on business processes, task constraints, and real-time operating status, enabling the embodied robot to adapt to multi-station, multi-equipment, and dynamic operating conditions in existing laboratories. After completing the actual physical operation, the robot feeds back the execution results to the digital twin and the compliance status, and links the perception data, state changes, action parameters, and execution results to form a traceability evidence chain. Therefore, this solution can not only reduce the risks of mishandling, misplacement, omission, and contact caused by repeated manual handling, opening, filling, preservation, and disposal of specimens, reagents, consumables, and hazardous materials, but also solve the problems of limited process coverage, insufficient linkage with quality rules, and lack of complete evidence records in the execution process of traditional automated equipment. This improves the adaptability, execution accuracy, process controllability, and quality traceability of standardized operations in medical laboratories. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of an embodied robot control method for standardized operations in a medical laboratory, as shown in one embodiment. Figure 2 This is a flowchart illustrating the steps involved in determining the candidate action set in one embodiment. Figure 3 This is a flowchart illustrating the steps involved in generating a target task sequence in one embodiment. Figure 4 This is a flowchart illustrating the steps involved in the control process performed by the embodied robot in one embodiment. Figure 5 This is a flowchart illustrating the steps involved in tracing the formation of a chain of evidence in one embodiment. Figure 6 This is a block diagram of an embodied robot control system for standardized operations in a medical laboratory, as shown in one embodiment. Detailed Implementation

[0017] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the specific embodiments of this application are described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0018] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0019] In this application, unless otherwise expressly specified and limited, the terms "connected" and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0020] It should be noted that when one element is considered to be "connected" to another element, it can be directly connected to the other element or connected to the other element through an intermediary element. Furthermore, in the following embodiments, "connection" should be understood as "electrical connection," "communication connection," etc., if there is transmission of electrical signals or data between the connected objects.

[0021] Medical laboratories undertake a large amount of specimen testing, retesting, preservation, reagent and consumable management, and hazardous materials disposal. Although the existing market offers automated pretreatment systems, posttreatment systems, fully automated laboratory production lines, single capping robots, sample sorting and storage systems, and reagent inventory management systems, they typically have the following limitations: First, they rely heavily on fixed tracks and modules, resulting in high construction costs and significant site modifications, making them difficult to adapt to small and medium-sized laboratories or scenarios where multiple brands of equipment coexist. Second, they primarily focus on specimen pretreatment or post-analysis storage, failing to cover the entire process including reagents and consumables, hazardous waste, cold chain, contact records, and review evidence. Third, there is insufficient linkage between automated equipment and LIS / LIMS / RIMS, quality systems, and clinical review rules. Fourth, they struggle to continuously learn from actual laboratory operating patterns and cannot automatically optimize based on events such as anomalies, retesting, stockouts, temperature control failures, and manual error correction.

[0022] In this embodiment, a medical laboratory refers to an operational environment that receives, pre-processes, tests, re-examines, preserves, and disposes of specimens, and manages the reagents and consumables required for testing, as well as hazardous materials generated during the process. A medical laboratory can be a hospital laboratory, an independent medical laboratory, or other testing facilities equipped with multiple testing devices and auxiliary workstations.

[0023] An embodied robot is a robot capable of sensing the state of target objects, equipment, and workspace within a medical laboratory, and performing physical actions such as movement, grasping, placing, opening and closing caps, adding, transferring, sealing, or handing over within the actual workspace according to generated task instructions. Embodied robots can include one or more of the following: mobile composite robots, fixed collaborative robotic arms, and embedded robotic arms positioned near specific work equipment.

[0024] The target objects of the operation include at least one of the following: specimens, reagents and consumables, and hazardous materials. Specimens may include blood, urine, body fluids, and other samples awaiting testing, contained in blood collection tubes, centrifuge tubes, sample cups, or other specimen containers; reagents and consumables may include reagent bottles, reagent kits, calibrators, quality control materials, reaction cups, pipette tips, and other testing consumables; hazardous materials may include infectious materials, chemically hazardous waste liquids, used sharp instruments, and other objects that need to be transported and disposed of according to preset safety conditions.

[0025] The operating equipment includes at least one of the following: analyzer, centrifuge, capping device, dispensing device, smart refrigerator, smart specimen storage cabinet, smart shelf, temperature-controlled transfer box, and hazardous waste temporary storage device. The operating space includes spatial locations such as specimen receipt station, centrifugation station, capping station, analyzer sample loading station, specimen preservation area, reagent library, reagent dispensing station, biosafety cabinet, and hazardous waste temporary storage area, as well as accessible paths connecting different spatial locations.

[0026] Standardized operations refer to operations performed on a target object in accordance with medical laboratory business processes, standard operating procedures, quality management rules, authority rules, and safety rules, with predetermined preconditions, actions to be performed, completion status, and abnormal handling methods. Standardized operations are not limited to a single action, but can include a complete work process consisting of multiple consecutive tasks.

[0027] This application relates to a method for controlling an embodied robot for standardized operations in medical laboratories, such as... Figure 1 As shown, the method includes the following steps S102 to S110. Wherein: S102. Acquire multimodal state data of the target work object, work equipment, work space and embodied robot in the medical laboratory, and update the digital twin corresponding to the target work object, work equipment, work space and embodied robot based on the multimodal state data.

[0028] Multimodal state data refers to data obtained from different types of sensors, operating equipment, and information systems, used to collectively characterize the state of the target work object and its operating environment. Multimodal state data can include at least two of the following: barcode recognition data, radio frequency identification data, visual image data, character recognition data, weight data, liquid level data, temperature and humidity data, position data, force data, collision detection data, and equipment operation data.

[0029] For specimens, multimodal status data may include specimen barcode or RFID tag, container appearance, cap type, specimen liquid level, specimen location, acquisition time, reception time, test items, centrifugation requirements, open status, on-machine status, and storage status.

[0030] For reagents and consumables, multimodal status data can include material identification, supplier, batch number, expiration date, packaging integrity, temperature control trajectory, inventory location, weight, remaining liquid level, open status, usage status, acceptance status, and calibration or performance verification status.

[0031] For hazardous materials, multimodal status data can include hazardous material identification, hazard type, hazard level, container type, packaging status, weight, current location, temperature control status, and contact records.

[0032] For operational equipment, multimodal status data can include whether the equipment is online, whether it is idle, current queue length, fault status, interface status, available time period, and current work object. For embodied robots, multimodal status data can include the robot's current position, movement status, remaining battery power, load status, end effector type, task occupancy status, and safety status. For the workspace, multimodal status data can include workstation occupancy status, ambient temperature and humidity, path accessibility, obstacle location, and area safety level.

[0033] Multimodal state data can also be obtained through the communication interfaces of LIS, LIMS, RIMS, instrument middleware, analyzers, centrifuges, refrigerators, and temperature control systems. LIS is a laboratory information system used to provide information related to specimens and testing items; LIMS is a laboratory information management system used to manage laboratory business processes, samples, and equipment information; RIMS is a reagent information management system used to manage reagent and consumable batch numbers, expiration dates, inventory, and usage records.

[0034] After acquiring multimodal state data, different modal data representing the same target operation object can be correlated and cross-validated. Data correlation refers to correlating data collected by different sensors with the same target operation object based on the acquisition time, acquisition location, vehicle where the target operation object is located, or business task.

[0035] A digital twin is a digital state mapping of a target work object, work equipment, work space, or embodied robot within an information processing system. A digital twin not only records the static attributes of the target object, but also its current state and historical state changes within a standardized work process.

[0036] Specifically, in an exemplary embodiment, the method further includes the process of establishing a digital twin. The process of establishing a digital twin includes: performing data association and state parsing on multimodal state data to obtain the object state of the target operation object, the operating state of the operation equipment and the embodied robot, and the environmental state of the work space; writing the object state, operating state, and environmental state into the state fields of the corresponding digital twin respectively; and associating the current operation state of the target operation object with the corresponding operation process nodes according to the business process corresponding to the target operation object to obtain the digital twin.

[0037] Through state analysis, the object state of the target task, the operational state of the task equipment and embodied robot, and the environmental state of the workspace can be obtained. The raw sensor data is thus transformed into structured state data that can be directly used for subsequent compliance status assessment and task graph construction.

[0038] Specifically, the parsed object status, running status, and environment status are written into the status field of the corresponding digital twin, and the current job status of the target job object is associated with the corresponding job process node according to the business process corresponding to the target job object.

[0039] A specimen digital twin may include fields such as specimen identification, test items, collection time, receipt time, container type, centrifugation requirements, centrifugation status, open status, machine status, re-examination status, storage location, storage period, hazard level, and disposal status.

[0040] The digital twin of reagents and consumables may include fields such as material identification, supplier information, batch number, expiration date, storage conditions, acceptance status, calibration or performance verification status, inventory location, remaining quantity, temperature control status, opening status, usage records, recall status, and scrap status.

[0041] A digital twin of hazardous materials can include fields such as hazardous material identification, hazard type, hazard level, container type, packaging status, temporary storage location, expiration date, disinfection status, contact records, and handover status.

[0042] Equipment digital twins can record equipment type, location, operating status, tasks to be performed, interface status, and available time periods. Spatial digital twins can record workstation location, traversable paths, path obstacles, area temperature and humidity, and area safety level. Robot digital twins can record robot location, remaining battery power, load status, end effector configuration, current task, and safety status.

[0043] When updating digital twins, the corresponding status fields of each digital twin can be updated based on the latest collected multimodal status data, and the associated status of the target work object in the business process nodes can be adjusted synchronously to ensure that the status of the digital twin is consistent with the real environment of the medical laboratory, providing an accurate data foundation for subsequent compliance judgments and task planning.

[0044] S104. Based on the digital twin and the preset medical laboratory operation rules, determine the current compliance status of the target operation object, and determine the set of candidate actions that the embodied robot is allowed to perform based on the current compliance status and the action authorization conditions associated with the current compliance status.

[0045] Among them, medical laboratory operation rules refer to the rules used to constrain the state transitions, operation sequence, and robot action permissions of the target operation objects. These rules may include laboratory standard operating procedures, quality management rules, object handling rules, equipment usage rules, biosafety rules, and operation permission rules.

[0046] For example, specimen handling rules can specify the sequence of specimen receipt, centrifugation, decapping, instrumentation, re-examination, recapping, preservation, and disposal of expired waste; reagent and consumable management rules can specify the conditions for reagent acceptance, new batch calibration or performance verification, warehousing, outbound, refilling, recall, and disposal; and hazardous materials disposal rules can specify the packaging methods, transfer routes, temporary storage areas, and access permissions corresponding to different hazard levels.

[0047] Medical laboratory operating rules can be derived from a pre-configured rule base or maintained by authorized personnel based on the laboratory's actual business processes. Changes to rules must be confirmed by the relevant authorized personnel to prevent unauthorized learning outcomes from directly altering mandatory rules related to quality and safety.

[0048] The compliance status can include at least one of the following: pending receipt, pending acceptance, pending calibration, pending performance verification, executable, near-expiration priority, abnormal suspension, disabled recall, pending manual review, saved, pending disposal, and archived.

[0049] Action authorization conditions refer to the conditions used to determine whether a robot action is allowed to be executed when the target work object is in a certain compliant state.

[0050] For example, when reagents are in the pending acceptance state, robots are allowed to perform the action of transporting them to the acceptance station, but are not allowed to perform the action of adding reagents to the analyzer's reagent compartment; when reagents are in the pending performance verification state, robots are allowed to perform the action of picking up and transporting them to the performance verification station, but are not allowed to perform the action of adding them as reagents for formal testing; when specimens are in the abnormal suspended state, robots are allowed to perform the action of isolating and transporting them or sending them to the manual verification station, but are not allowed to perform the action of automatically loading them onto the machine.

[0051] For hazardous materials, grabbing, transferring, or temporarily storing actions are only permitted when the hazard level, container type, packaging status, planned transfer route, and access permissions all meet the corresponding rules.

[0052] The preset set of robot actions can include actions such as recognition, grasping, placing, transporting, opening and closing caps, adding fillings, sealing, disinfection, transfer, warehousing, warehousing, and handover. By filtering the set of candidate actions that the embodied robot can perform, actions that do not meet the current compliance status and action authorization conditions are eliminated, narrowing the range of action options. This ensures that the operation complies with the quality and safety specifications of the medical laboratory and avoids ineffective action planning that reduces operational efficiency.

[0053] S106. Based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, as well as the candidate action set, construct an executable task graph and generate a target task sequence according to the executable task graph.

[0054] Among them, a business process refers to a process that is pre-set for a certain type of target operation object and consists of multiple standardized operation stages.

[0055] The specimen handling process may include stages such as receipt, identity verification, centrifugation, removal of cap, visual inspection, machine testing, re-examination or retesting, capping, preservation, and waste disposal.

[0056] The reagent and consumable business process may include stages such as signing for receipt, identity verification, acceptance, calibration or performance verification, warehousing, outbound, opening boxes, taking bottles, opening caps, transfer, filling, returning to location, inventory, recall and disposal.

[0057] The hazardous materials business process may include stages such as identification, classification, packaging, transfer, temporary storage, disinfection, handover and disposal upon expiration.

[0058] Task constraints are the conditions that determine whether a task node can be executed at present. They can include prerequisite task constraints, compliance status constraints, task time limit constraints, availability constraints of operating equipment, availability constraints of robot, time limit constraints for saving target work objects, temperature control constraints, path access constraints, and safety level constraints.

[0059] The preceding task constraint indicates that the preceding task on which the task node depends has been completed; the compliance status constraint indicates that the current compliance status of the target operation object allows the corresponding action to be executed; the task time limit constraint indicates that the task needs to be completed within a preset time; the equipment availability constraint indicates that the corresponding analyzer, centrifuge, storage cabinet or filling equipment is in an available state; the robot availability constraint indicates that the robot is not currently occupied by other tasks, the remaining power meets the requirements and is configured with a suitable end effector.

[0060] The executable task graph is determined by the executable status of task nodes. It is obtained by filtering nodes in the business process, combined with the current digital twin status and task constraints. It includes executable task nodes that currently meet all constraints, and each executable task node corresponds to one of the allowed actions in the candidate action set.

[0061] Select the executable task node from the executable task graph, and generate the target task sequence according to the preset task priority and the execution order represented by the task association edge. The target task sequence contains multiple specific robot action tasks to be executed in sequence. Each task will specify the work object, target position, type of action to be performed, and corresponding work requirements to ensure that the task execution order complies with business processes and safety specifications.

[0062] S108. Control the embodied robot to perform standardized operations for the target object in different workstations according to the target task sequence, and obtain the execution results.

[0063] When performing each specific task, the embodied robot collects real-time status data of the operation process through its own multi-sensor system, updates the status fields of the corresponding digital twin synchronously, and feeds back the execution status to the control module in real time to ensure that each action meets the operation specifications and safety requirements.

[0064] S110. Update the digital twin and current compliance status based on the execution results, and associate the multimodal status data, compliance status changes, action parameters and execution results to form a traceability evidence chain for standardized operations.

[0065] Specifically, once the embodied robot completes all its pre-defined tasks or interrupts a task due to an anomaly, it updates the status fields of the corresponding digital twin based on the actual execution results. Simultaneously, it reassesses the compliance status of the target task object and adjusts its association with business process nodes. The traceability chain comprehensively records key information from data collection, compliance assessment, task planning to action execution, including the collection time, collection equipment, and raw values ​​of all multimodal status data, the process of compliance status changes, and the execution parameters and results of each action. This not only meets the medical laboratory's quality management requirements for traceability throughout the entire testing process but also facilitates rapid identification of the cause when anomalies occur, providing complete data support for quality review and problem rectification.

[0066] The aforementioned embodied robot control method for standardized operations in medical laboratories maps the target work object, work equipment, work space, and the multimodal state of the embodied robot to a digital twin. This enables the system to grasp the actual operational stage of specimens, reagents, consumables, or hazardous materials, as well as the status of surrounding equipment and the robot. Furthermore, it utilizes medical laboratory operating rules to determine the current compliance status and implements authorization gating for robot actions, ensuring that the robot only receives the corresponding candidate action when the object state, quality rules, and operating permissions all meet the requirements. Based on this, an executable task graph is constructed and a target task sequence is generated according to business processes, task constraints, and real-time operating status, enabling the embodied robot to adapt to multi-station, multi-equipment, and dynamic operating conditions in existing laboratories. After completing the actual physical operation, the robot feeds back the execution results to the digital twin and the compliance status, and links the perceived data, state changes, action parameters, and execution results to form a traceability evidence chain. Therefore, this solution can not only reduce the risks of mishandling, misplacement, omission, and contact caused by repeated manual handling, opening, filling, preservation, and disposal of specimens, reagents, consumables, and hazardous materials, but also solve the problems of limited process coverage, insufficient linkage with quality rules, and lack of complete evidence records in the execution process of traditional automated equipment. This improves the adaptability, execution accuracy, process controllability, and quality traceability of standardized operations in medical laboratories.

[0067] In one exemplary embodiment, such as Figure 2 As shown, based on the digital twin and preset medical laboratory operating rules, the current compliance status of the target operating object is determined. Based on the current compliance status and the associated action authorization conditions, a set of candidate actions that the embodied robot is allowed to perform is determined, including: S1041. Convert medical laboratory operation rules into status recognition conditions and action authorization conditions.

[0068] Among them, the status recognition condition refers to the condition for determining the current compliance status of the target operation object based on the status field in the digital twin.

[0069] For example, for reagents and consumables, when the packaging is intact and the temperature control trajectory meets the requirements but acceptance has not yet been completed, it can be identified as a pending acceptance state; when a new batch of reagents has passed acceptance but calibration or performance verification has not yet been completed, it can be identified as a pending calibration state or a pending performance verification state; when acceptance, calibration or performance verification all meet the requirements, it can be identified as an executable state; when a recall event exists, it can be identified as a disabled recall state.

[0070] For specimens, when the specimen identification information is inconsistent, the container is damaged, or the time from collection to receipt exceeds the allowable time limit, it can be determined as an abnormal pending state or a state awaiting manual review; when the specimen has been tested and preserved as required, it can be determined as a preserved state; when the preservation period is reached, it can be determined as a state awaiting disposal.

[0071] Action authorization conditions refer to the conditions used to determine whether a robot action is allowed to be executed when the target work object is in a certain compliant state.

[0072] For example, when reagents are in the pending acceptance state, robots are allowed to perform the action of transporting them to the acceptance station, but are not allowed to perform the action of adding reagents to the analyzer's reagent compartment; when reagents are in the pending performance verification state, robots are allowed to perform the action of picking up and transporting them to the performance verification station, but are not allowed to perform the action of adding them as reagents for formal testing; when specimens are in the abnormal suspended state, robots are allowed to perform the action of isolating and transporting them or sending them to the manual verification station, but are not allowed to perform the action of automatically loading them onto the machine.

[0073] For hazardous materials, grabbing, transferring, or temporarily storing actions are only permitted when the hazard level, container type, packaging status, planned transfer route, and access permissions all meet the corresponding rules.

[0074] S1042. Match the object status of the target operation object in the digital twin with the status identification conditions to determine the current compliance status of the target operation object.

[0075] Specifically, based on the object type of the target task object and its currently associated task process node, target status identification conditions corresponding to the object type and task process node are selected from the pre-established status identification conditions.

[0076] Read the status fields in the digital twin that correspond to the target status identification conditions, and match the current value of each status field with the target status identification conditions one by one. When a status field meets the status identification conditions corresponding to a certain compliance status, the compliance status is determined as the current compliance status of the target work object; when the object status has inconsistent identification results, abnormal packaging or containers, abnormal temperature control, or other situations that do not meet the normal circulation conditions, the current compliance status is determined as an abnormal suspended status or a status pending manual review.

[0077] For example, for reagents and consumables, when they have been signed for but not yet accepted, they are identified as pending acceptance based on the acceptance status field; when reagents and consumables have been accepted but new batches of reagents have not yet been calibrated or verified, they are identified as pending calibration or performance verification; when acceptance, calibration or performance verification has been completed and there are no abnormal temperature control, recall or disabling information, they are identified as executable.

[0078] For specimens, their current compliance status can be determined based on fields such as identity verification results, collection and receipt times, container type, centrifugation status, open status, machine-mounted status, and storage status. For example, specimens that have passed identity verification and meet the conditions for subsequent processing can enter the executable state; specimens with inconsistent identity information, damaged containers, or that do not meet the conditions for subsequent processing can enter the abnormal suspension state or the state awaiting manual review. The centrifugation, open status, machine-mounted status, review status, and storage status generated during specimen processing are all written into the specimen's digital twin for subsequent status matching.

[0079] S1043. Based on the current compliance status, query the associated action authorization conditions to obtain the permitted and prohibited actions under the current compliance status.

[0080] Specifically, a pre-established association between different compliance statuses and action authorization conditions is established. Each action authorization condition includes at least the corresponding compliance status, the applicable target object type, and the robot actions that are allowed or prohibited. These robot actions may include at least one of the following: grasping, placing, transferring, centrifugal loading, removing caps, replacing caps, loading, warehousing, warehousing, opening boxes, adding fillings, sealing, temporary storage, disinfection, and handover.

[0081] Based on the current compliance status determined in step S1042, query the action authorization conditions associated with the current compliance status, and match each robot action in the preset robot action set with the action authorization conditions. Robot actions that meet the permitted conditions corresponding to the current compliance status are determined as permitted actions; robot actions that are explicitly prohibited by the current compliance status or do not meet the corresponding authorization conditions are determined as prohibited actions.

[0082] For example, when reagents and consumables are in a state awaiting acceptance, the embodied robot is allowed to perform actions such as transferring the reagents and consumables to the acceptance station, isolating and storing them, or verifying information, but it is prohibited to perform actions such as formal use or adding reagents to the analyzer; when reagents and consumables are in a state awaiting calibration or performance verification, it is allowed to perform actions such as transferring them to the calibration or performance verification station, but it is prohibited to perform actions such as formal testing and adding reagents before calibration or performance verification is completed.

[0083] When the specimen is in an executable state and the current process node is waiting to be centrifuged, waiting to be decapped, or waiting to be loaded onto the machine, centrifugation loading, decapping, or loading and transfer actions can be performed respectively. When the specimen is in an abnormal suspended state or waiting for manual review, actions such as isolation transfer, transfer to manual review station, or safe storage can be performed, while automatic centrifugation, automatic decapping, or automatic loading actions are prohibited.

[0084] For hazardous materials, packaging, transfer, temporary storage, or handover actions are only permitted when the hazard level, container type, packaging status, transfer route, disinfection requirements, and access permissions meet the corresponding authorization conditions. When there is leakage, damage, temperature control failure, or path obstruction, continuing the transfer or handover will be deemed a prohibited action, while safety actions such as stopping operations, entering a safe state, and generating emergency response tasks will be permitted.

[0085] S1044. Eliminate prohibited actions from the preset robot action set and determine the remaining allowed actions as the candidate action set.

[0086] Specifically, by excluding prohibited actions from the preset set of robot actions and retaining only permitted actions that meet the authorization requirements of the current compliance status, non-compliant operations can be avoided at the source, strictly adhering to the quality and safety management requirements of medical laboratories. This also narrows the search scope for subsequent task planning, avoids generating invalid tasks that do not meet the rule requirements, and further improves the efficiency and accuracy of task planning.

[0087] The candidate action set does not mean that each action will be executed immediately, but rather that these actions have passed compliance gating and can be further judged for their executability in the subsequent task graph construction process, taking into account equipment status, robot status, and task constraints.

[0088] In one exemplary embodiment, such as Figure 3 As shown, based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, and the set of candidate actions, an executable task graph is constructed, and a target task sequence is generated according to the executable task graph, including; S1061. Decompose the business process corresponding to the target job object into multiple task nodes, and configure task attributes for each task node; wherein, the task attributes include target job object, previous state, target state, target workstation, candidate action, completion time limit and exception handling conditions.

[0089] The business process corresponding to the target task is broken down into multiple task nodes. Each task node corresponds to a work unit that can be completed by a robot, work equipment, or human reviewer.

[0090] The task attributes of a task node can include the target work object, prerequisite status, target status, target workstation, target action, candidate actions, completion time limit, and exception handling conditions. The target work object is the specimen, reagent / consumable, or hazardous material entity to be processed in the current task. The prerequisite status refers to the compliance status of the target work object that must be met to enter the current task node. The target status refers to the compliance status that the target work object should reach upon successful completion of the current task node. The target workstation specifies the physical location where the task is executed. The candidate actions are the range of actions that the current task node can perform. The completion time limit specifies the upper limit of the time required to complete the task. The exception handling conditions specify the handling methods when exceptions occur during task execution.

[0091] For example, the preceding state of the "Specimen Decapping Task Node" can be "Centrifugation Completed", the target state can be "Decapping Completed", the target workstation can be "Decapping Workstation", and the target action can be "Grab the test tube and perform rotation and lifting".

[0092] S1062. Establish task association edges based on the execution sequence and compliance status transition relationships between each task node to obtain the initial task graph.

[0093] In this context, the execution sequence refers to the order of operations between two tasks. For example, a specimen will only proceed to the decapping task after centrifugation is completed; reagents will only proceed to the formal outbound and dispensing tasks after acceptance is completed and calibration or performance verification requirements are met.

[0094] The compliance status transition relationship indicates that the actual completion status of the previous task can trigger the target operation object to transition from the current compliance status to the next compliance status, and enable the subsequent task to be activated.

[0095] An initial task graph is formed by task nodes and task-related edges. The initial task graph is used to describe the logical relationships between tasks in a complete business process, but not all task nodes in it are necessarily executable at the current moment.

[0096] S1063. Based on the operating status of the working equipment and the embodied robot, the set of candidate actions, and the task constraints, determine the executable status of each task node in the initial task diagram to form an executable task diagram.

[0097] Specifically, for each task node in the initial task graph, it is first determined whether the corresponding target action belongs to the candidate action set determined in step S104. If the target action does not belong to the candidate action set, it means that the action has not been authorized by the current compliance status, and the corresponding task node is not set to an executable state. Specifically, in an exemplary embodiment, the executable state of each task node in the initial task graph is determined according to the operating state of the working equipment and the embodied robot, the candidate action set, and the task constraints, forming an executable task graph, including: for each task node in the initial task graph, determining whether the target action corresponding to the task node belongs to the candidate action set, and determining whether the working equipment and embodied robot required to perform the target action are available according to the operating state of the working equipment and the embodied robot; when the target action belongs to the candidate action set, the required working equipment and embodied robot are available, and the task node meets the corresponding task constraints, the task node is determined to be in an executable state; otherwise, the task node is determined to be in a waiting state or a blocked state; the task nodes and task-related edges in the initial task graph are updated according to the executable state of each task node to form an executable task graph.

[0098] For example, the sample loading task requires the corresponding analyzer to be online and have an available sample loading location, and the robot to be in a schedulable state; the centrifugation task requires the centrifuge to be in a usable state and to have a centrifugation adapter that meets the sample type and centrifugation conditions; the reagent dispensing task requires the reagent compartment of the target analyzer to be accessible, and the robot to be configured with a reagent bottle gripping end or a dispensing end.

[0099] If the prerequisite task has not been completed, the equipment is currently occupied, or the robot is temporarily unavailable, but the conditions can be met after waiting, the task node can be set to a waiting state. If the target action is prohibited, the target work object is in an abnormal suspended state, the work equipment is malfunctioning, or the path is impassable, the task node can be set to a blocked state.

[0100] Subsequently, the task nodes and task-related edges in the initial task graph are updated according to the executable status of each task node. For example, the task-related edge between a blocked task node and a subsequent task node can be set to temporarily impassable; once the preceding task is completed and the equipment is restored to usable status, the corresponding edge can be reset to passable status, thereby forming an executable task graph that reflects the current actual working conditions.

[0101] S1064. Extract the target task sequence from the executable task graph according to the preset task priority and the execution order represented by the task association edges.

[0102] Select the executable task nodes from the executable task graph and generate the target task sequence according to the preset task priority and the execution order represented by the task association edges.

[0103] Task priorities can be determined based on factors such as the time limit of the testing item, whether it is an emergency specimen, the length of the centrifugation queue, the availability of the analyzer, the specimen preservation time limit, the retesting or re-examination requirements, the remaining reagents, the temperature control status, the risk level of the target object, and the remaining battery power of the robot.

[0104] For example, when there are both routine specimen transport tasks and emergency specimen loading tasks with short testing time limits, the emergency specimen loading task can be generated first; when the robot's remaining battery power is lower than the power required to complete long-distance transport, the transport task can be suspended and the charging task or the task performed by other robots can be added to the target task sequence.

[0105] The target task sequence can include multiple tasks executed consecutively by the same robot, or different tasks can be assigned to multiple robots or fixed robotic arms for collaborative execution.

[0106] If, during the execution of the target task sequence, it is detected that the working equipment is unavailable, the path is blocked, the target working object is in an abnormal state, the robot has insufficient power, the action execution fails, or the task times out, the executable status of the corresponding task node or task-related edge can be updated, and the target task sequence can be regenerated based on the updated executable task graph.

[0107] For example, if the scheduled analyzer malfunctions, the original on-machine task node can be set to a blocked state, and a task to transfer to another analyzer can be generated according to the instrument switching rules; if the container is found to be damaged during the sample capping process, the subsequent on-machine tasks can be terminated, and an isolation transfer and manual verification task can be generated.

[0108] By dynamically updating the executable task graph, fixed task sequences can be prevented from continuing to execute mechanically after changes in device or object states.

[0109] In one specimen processing embodiment, a robot or specimen receiving station reads the specimen barcode or RFID tag and obtains the test items, collection time, container type, priority, whether it is an emergency, whether centrifugation is required, and risk information via LIS. The system establishes or updates a digital twin of the specimen and determines whether the specimen meets the conditions for entering the subsequent process based on the time difference between the collection time and the receiving time, the time limit of the test items, and the container type.

[0110] Specimens requiring centrifugation are entered into the centrifugation task queue. A centrifugation loading plan is generated based on specimen type, project requirements, centrifugation speed, centrifugation time, centrifugation temperature, and balancing rules. The robot places the specimen into the centrifuge adapter or centrifuge cup. Centrifugation is allowed to start after the centrifugation balancing meets the requirements; if the balancing does not meet the requirements, a compensation balancing task or a manual verification task can be generated.

[0111] After centrifugation, determine whether the specimen needs to be decapped, sent to the designated analyzer, or moved to the dispensing station or buffer station according to the task diagram.

[0112] During cap removal, the cap type is determined by visual recognition, and the rotational torque and upward pulling force are adjusted based on force feedback. After cap removal, information on specimen liquid level, clots, or abnormal appearance can be obtained, and the recognition results are written into the specimen's digital twin. The appearance recognition here is used to determine whether the specimen meets the conditions for subsequent standardized procedures, rather than directly making a disease diagnosis.

[0113] During the on-machine task, the specimen is sent to the corresponding analyzer or buffer area according to the analyzer's operating status and the combination of test items.

[0114] When the test results trigger preset rules for retesting, dilution, re-examination, transfer, or critical value verification, the robot receives the corresponding event from the LIS or instrument middleware, relocates the specimen, and adds it to the corresponding task queue. If the corresponding rule requires manual confirmation, an authorized personnel verification task is generated, and subsequent tasks are executed only after the confirmation result is received. After completing the test, the robot can perform tasks such as resealing, saving, partitioned storage, and waste disposal.

[0115] It should be noted that the terms "retesting," "re-examination," "dilution," "transfer testing," or "instrument switching" in this application are used to determine subsequent work tasks for the specimen and do not replace medical diagnosis by physicians or laboratory professionals. Recommendations involving professional judgment should be confirmed by authorized personnel before implementation.

[0116] In one embodiment of reagent and consumable management, a robot or reagent receiving station collects supplier information, batch number, expiration date, packaging integrity, temperature control trajectory, and registration or filing information of reagents and consumables, and writes the corresponding information into a digital twin of the reagents and consumables.

[0117] After the reagent is received, its compliance status is set to "pending acceptance." Upon successful acceptance, if the reagent is from a new batch or affects testing performance, its compliance status is changed to "pending calibration" or "pending performance verification." Until calibration or performance verification is completed, the robot is prohibited from using the reagent for formal testing or dispensing under the authorized action conditions.

[0118] Upon receipt, the corresponding temperature zone and compartment are determined based on the storage conditions of the reagents and consumables. The robot is then controlled to place the reagents and consumables in the corresponding positions of the smart refrigerator or smart shelf, and the actual inventory location is written into the digital twin of the reagents and consumables.

[0119] When issuing or refilling, a retrieval task is generated based on the testing plan, remaining inventory, first-in-first-out (FIFO) or near-expiry priority rules, and batch number verification status. The robot can complete the tasks of opening boxes, retrieving bottles, opening caps, transferring, refilling, returning to its original location, and updating remaining inventory according to the target task sequence.

[0120] For example, when the analyzer needs reagent replenishment, the system generates a retrieval task based on the required reagent type, the location of the target reagent bottle, the remaining liquid level, and the batch status. The robot moves to the reagent storage location, uses its reagent bottle gripping end to grab the reagent bottle, transports it to the analyzer's reagent compartment, and performs pouring, pumping, or injecting actions according to the filling requirements. After filling, the remaining reagent level is updated based on changes in weight or liquid level.

[0121] When inventory falls below a preset threshold, reagents are nearing their expiration date, temperature control is abnormal, consumption is abnormal, or a recall event occurs, replenishment, transfer, disabling, recall, or scrapping tasks can be generated, and the compliance status of the digital twin of reagents and consumables can be updated according to the corresponding event.

[0122] For target work objects with risks of infection, chemical hazards, or sharp instruments, establish a digital twin of the hazardous materials and record the hazard type, hazard level, container type, packaging status, current location, temporary storage location, expiration time, and contact records.

[0123] Based on the hazard level, container type, and packaging status of the hazardous materials, determine the corresponding packaging conditions, transfer routes, temporary storage locations, disinfection conditions, and access authorization conditions. Only when the above conditions are met will the corresponding grabbing, packaging, transfer, temporary storage, or handover actions be added to the candidate action set.

[0124] When robots perform hazardous materials transfer and packaging, they should move along restricted paths and avoid entering areas that do not meet the corresponding safety level. Each contact should be documented with robot identification, personnel identification, operation time, operation location, hazardous materials identification, disinfection method, and handover results.

[0125] When hazardous materials reach a preset capacity threshold or expire, a packaging and handover task is generated; if leakage, damage, temperature control failure, or path obstruction occurs, the current task is stopped, the robot is controlled to enter a preset safety state, and an emergency response task is generated.

[0126] Preset safety states can include stopping movement, holding or lowering the current target work object, shutting down relevant end effectors, exiting to a safe area, and sending abnormal information to the monitoring terminal. The specific safety actions to be taken can be pre-configured based on the type of hazardous material and the current work location.

[0127] In one exemplary embodiment, such as Figure 4 As shown, the robot is controlled to perform standardized tasks for the target object at different workstations according to the target task sequence, and the execution results are obtained, including: S1081. Based on the target work object, target action and target workstation in the target task sequence, determine the target execution device of the embodied robot, and determine the target end effector from multiple replaceable end effectors.

[0128] When the target task involves transfer between different workstations, a mobile composite robot can be selected; when the target task involves performing precise grasping, opening, or dispensing at a fixed workstation, a fixed collaborative robotic arm or an embedded robotic hand can be selected; when a task simultaneously involves cross-workstation movement and precise operations, it can be performed by a mobile robot carrying a robotic arm, or by a mobile robot and a fixed robotic arm working together. Preferably, standardized operations targeting the target work object are performed at one or more workstations.

[0129] A target end effector is an execution component installed at the end of a robot arm that directly acts on the target object. It can include test tube gripping end, cap opening end, cap closing end, reagent bottle gripping end, box opening end, dispensing end, and hazardous waste sealing end, etc.

[0130] Based on the type of the target work object, the container structure, and the target action, determine the appropriate target end effector from multiple replaceable end effectors.

[0131] For example, for test tube transport tasks, select the test tube gripping end; for test tube cap removal tasks, select the cap opening end; for reagent bottle retrieval tasks, select the reagent bottle gripping end; and for hazardous material packaging tasks, select the hazardous waste packaging end.

[0132] S1082. Generate the initial motion parameters of the target execution device and the target end effector based on the structure and current spatial location of the target work object.

[0133] Specifically, based on the location information, container dimensions, and current posture recorded in the digital twin of the target work object, and combined with the equipment coordinate system of the work station, the movement path nodes of the target execution device, the range of joint rotation of the robotic arm, and the initial gripping position, clamping force, and action angle of the target end effector are calculated. This ensures that the motion parameters are adapted to the state of the target work object in the current scenario, avoiding gripping failure or object damage caused by fixed parameters.

[0134] For test tubes, structural information may include test tube specifications, cap type, tube rack hole position, and test tube orientation; initial motion parameters may include gripping position, clamping opening, gripping force, rotational torque, pull-up distance, and moving speed.

[0135] For reagent bottles, structural information may include bottle type, cap type, bottle center of gravity, remaining liquid level, and target instrument reagent compartment location; initial action parameters may include gripping position, clamping force, cap opening direction, tilting angle, pump suction or injection volume.

[0136] For hazardous material containers, structural information may include container dimensions, weight, encapsulation interface, and permissible orientation; initial action parameters may include clamping position, moving speed, safety distance, and encapsulation action parameters.

[0137] S1083. Execute the corresponding target task according to the initial action parameters, and obtain the execution feedback results during the execution process; wherein, the execution feedback results include at least one of the following: visual feedback, position feedback, force feedback and collision feedback.

[0138] Specifically, the corresponding target task is executed according to the initial action parameters, and at least one of visual feedback, positional feedback, force feedback and collision feedback is obtained during the execution process.

[0139] Visual feedback can be used to determine the relative position between the target object and the end effector, and to identify gripping offset or container posture; position feedback can be used to determine the actual position of the robot, robotic arm, and end effector; force feedback can be used to determine whether the gripping is stable, whether the opening torque is appropriate, or whether jamming has occurred; collision feedback can be used to identify abnormal contact between the robot or robotic arm and surrounding equipment.

[0140] S1084. Adjust the initial action parameters based on the execution feedback results, and after completing the standardized operation of the current target task, jump to the next task in the target task sequence until all operations in the target task sequence are completed, and obtain the state of the target operation object after execution.

[0141] Specifically, the initial motion parameters are adjusted based on the execution feedback. For example, when visual feedback indicates a shift in the gripping position, the position of the robotic arm's end effector is adjusted; when force feedback indicates insufficient opening torque, the rotational torque is increased within the allowable range; when force feedback indicates excessive gripping force, the gripping force is reduced; and when collision detection is triggered, the current motion is stopped and the local motion path is replanned.

[0142] After completing the current target task, the robot moves on to the next task in the target task sequence until all tasks are completed. This execution method enables the task sequence generated by the task orchestration layer to be transformed into continuous actions of the embodied robot in a real environment.

[0143] S1085. Compare the status of the target job object after execution with the expected completion status of the target task to determine the execution result: successful execution, failed execution, or abnormal execution.

[0144] After completing the target task, obtain the status of the target job object after execution and compare it with the expected completion status corresponding to the target task.

[0145] For example, for the task of removing the cap, the expected completion status is that the test tube cap has been separated from the test tube and the test tube remains stable; for the task of loading the instrument, the expected completion status is that the specimen has been placed in the designated location of the target analyzer; for the task of adding reagents, the expected completion status is that the amount of reagent in the target reagent compartment has reached the required level and the status of the reagent bottle has been updated; for the task of transferring hazardous materials, the expected completion status is that the hazardous materials have been placed in the designated temporary storage location and the handover record has been completed.

[0146] When the status of the target task object after execution meets the expected completion status, it is determined as successful execution; when the action is not completed and cannot continue to be executed, it is determined as failed execution; when the action is completed but positional deviation, container damage, abnormal force, or other uncertain states are detected, it is determined as abnormal execution.

[0147] The execution results are used for subsequent digital twin updates, compliance status transitions, and evidence chain recording, thereby creating a closed loop between the action execution steps and subsequent status management steps.

[0148] In one exemplary embodiment, such as Figure 5 As shown, the digital twin and current compliance status are updated based on the execution results, and the multimodal status data, compliance status changes, action parameters, and execution results are correlated to form a traceability evidence chain for standardized operations, including: S1101. Determine the actual task completion status corresponding to the target task object based on the execution results.

[0149] The actual task completion status can include tasks completed, tasks not completed, tasks partially completed, objects in abnormal status, or requiring manual confirmation. Specifically, successful execution corresponds to a completed task status, execution failure corresponds to a not completed task status, and execution exception corresponds to an objects in abnormal status or requiring manual confirmation.

[0150] S1102. Write the actual task completion status into the digital twin.

[0151] The actual completion status is written into the digital twin corresponding to the target task object, and the operation equipment, operation space and embodied robot digital twin related to the current target task are updated simultaneously.

[0152] For example, after the robot completes the specimen loading task, the loading status in the specimen digital twin is updated to "loaded", the current working object in the analyzer digital twin is updated to the specimen, and the current task in the robot digital twin is updated to "completed".

[0153] Once the reagent filling task is completed, update the remaining quantity, usage record, and current location in the reagent consumable digital twin, and simultaneously update the analyzer reagent compartment status and robot task status.

[0154] S1103. Match the actual task completion status with the status transition conditions corresponding to the current compliance status.

[0155] Among them, the state transition conditions are conditions that are pre-set in association with the current compliance status and the target task, used to determine whether the target operation object can enter the next compliance status. These conditions can include at least the target task completion conditions, the target operation object status conditions, the data integrity conditions, and the anomaly exclusion conditions.

[0156] Specifically, based on the current compliance status and the current target task, the corresponding target state transition condition is determined from the preset state transition conditions, and the state field corresponding to the target state transition condition is read from the updated digital twin; the actual task completion status and the current value of the state field are compared with the target state transition condition item by item to determine whether the execution of the current target task enables the target job object to meet the requirements for entering the next job stage.

[0157] For example, for specimen decapping tasks, state transition conditions may include successful decapping, separation of the cap from the specimen container, no damage to the specimen container, and consistency of specimen identification information. When all of the above conditions are met, it indicates that the specimen can move from the current state of awaiting decapping to the state of awaiting instrumentation. For reagent acceptance tasks, state transition conditions may include completion of the acceptance task, intact packaging, compliance of the temperature control trajectory, and verification of batch number and expiration date information. If the reagent is a new batch reagent requiring calibration or performance verification, the next compliance state is the state of awaiting calibration or performance verification, rather than directly entering the executable state.

[0158] By matching the actual task completion status with the state transition conditions, we can avoid directly changing the compliance status simply because the robot's action has ended, ensuring that changes in the compliance status are constrained by both the action execution result and the actual state of the target work object.

[0159] S1104. When the state transition conditions are met, the current compliance state is transitioned to the corresponding next compliance state.

[0160] Specifically, when the actual task completion status and the relevant status fields in the digital twin both meet the target status transition conditions, the next compliance status of the target operation object is determined based on the pre-established correspondence between the current compliance status, the target task and the next compliance status, and the current compliance status recorded in the digital twin is updated to the next compliance status.

[0161] At the same time, the previous compliance status, the next compliance status, the target task that triggered the status transition, the status transition time, and the basis for the status transition are written into the digital twin so that the updated compliance status can be used to determine the candidate actions that can be allowed to be executed and to activate the subsequent task nodes in the executable task graph.

[0162] For example, after a specimen has been centrifuged and the centrifugation results meet the requirements, it can be changed from the state of waiting to be centrifuged to the state of waiting to be capped or waiting to be loaded into the machine; after a specimen has been tested and capped back for storage, it can be changed to the state of being stored; when a reagent has been accepted but still needs performance verification, it can be changed from the state of waiting to be accepted to the state of waiting to be verified; when a reagent has been calibrated or verified for performance and there are no abnormal temperature control, recall or disabling information, it can be changed to the state of being executable.

[0163] The next compliance status is not necessarily a fixed subsequent status in the business process, but is determined based on the actual completion of the target task and the corresponding business process. For example, specimens that do not need to be capped can directly enter the "ready to be used" state after centrifugation, while specimens that need to be aliquoted can enter the "ready to be aliquoted" state. Thus, the compliance status transition can be consistent with the specific business process of the target operation.

[0164] S1105. When the state transition conditions are not met or the execution result is execution failure or execution exception, the target job object is converted to an abnormal suspended state or a state awaiting manual review, and an exception handling task is generated.

[0165] Among them, the abnormal suspension status means that the normal business process of the target operation object is temporarily suspended and subsequent actions that may amplify errors or safety risks are prohibited; the status pending manual review means that the current abnormality cannot be automatically determined based on the existing status data, and the subsequent operation needs to be determined by authorized personnel after confirming the abnormality information.

[0166] Specifically, when the actual task completion status does not meet the corresponding status transition conditions, the system determines whether the target job object can be restored to normal operation through automatic handling, based on the unmet status fields, execution feedback results, and exception types.

[0167] When an anomaly has a clearly defined automatic handling method, the target work object is converted to an anomaly suspended state, and a corresponding anomaly handling task is generated. The anomaly handling task may include at least one of the following: re-identification, re-capture, relocation, compensation and balancing, retransfer, isolation and storage, switching work equipment, changing the transfer route, or entering a safe state.

[0168] For example, when a centrifugation task fails to be executed due to insufficient balancing, a compensation balancing task can be generated; when the target analyzer is temporarily unavailable, a waiting task or a transfer task to switch to another available analyzer can be generated; when a path obstacle is detected during robot transfer, a task to replan the path can be generated.

[0169] When an anomalies involve inconsistencies in object identity, damaged specimen containers, abnormal reagent temperature control, reagent recall, hazardous material leakage, or when the cause of the anomaly cannot be automatically determined, the target work object is converted to a state awaiting manual review, and a manual review task is generated that includes the cause of the anomaly, relevant sensing data, execution feedback results, current location, and suggested handling methods.

[0170] In the event of hazardous material leakage, damage, temperature control failure, or obstruction of the transfer route, the current task should be stopped, and the embodied robot should be controlled to enter a preset safety state to prevent the continued execution of transfer or handover actions from increasing safety risks.

[0171] Before the abnormal suspended state or the state awaiting manual review is lifted, update the subsequent task nodes corresponding to the target operation object in the executable task graph, so that they are in a waiting state or a blocked state, thereby preventing the abnormal object from continuing to enter the subsequent centrifugation, cap removal, machine loading, refilling or disposal process.

[0172] S1106. Based on the updated content, compliance status transition information and action parameters of the digital twin, generate an evidence chain record corresponding to the target operation object, and establish the association relationship between multiple evidence chain records according to the execution order of each target task in the target task sequence to form a traceability evidence chain for standardized operations.

[0173] The updated content of the digital twin refers to the changes in the object state, equipment state, spatial state, and robot state before and after the execution of the current target task; the compliance state transition information includes at least the previous compliance state, the next compliance state, the state transition conditions, and their matching results; the motion parameters refer to the gripping position, gripping force, rotational torque, moving speed, opening angle, filling amount, or other control parameters corresponding to the target action used by the embodied robot when executing the current target task.

[0174] For each target task, a task identifier is assigned, and the task identifier is associated with the target operation object identifier, rule basis, standard operating procedure basis, robot identifier, end effector identifier, operator identifier, operation time, operation location, multimodal state data, motion parameters, execution result, digital twin update content, compliance status transition information, anomaly type and human intervention information to generate a corresponding evidence chain record.

[0175] According to the execution order of each objective task in the objective task sequence, the current evidence chain record is associated with the evidence chain record corresponding to the previous objective task. For example, the record identifier or integrity verification information of the previous evidence chain record can be written into the current evidence chain record, so that the evidence chain records corresponding to different objective tasks such as signing, centrifugation, removing the cap, loading, review, recapping, preservation and disposal are connected sequentially.

[0176] The chain of evidence can be queried according to the target task object, task identifier, rule basis, or abnormal event.

[0177] For example, when investigating a quality incident involving a specimen, the entire chain of evidence records from receipt, centrifugation, decapping, instrumentation, re-inspection, preservation to disposal can be retrieved based on the specimen identification; when recalling a batch of reagents, the records of acceptance, warehousing, outbound, dispensing, and use of that batch of reagents can be retrieved; and when inspecting the handover of hazardous materials, the records of hazardous material transfer, disinfection, contact, and handover can be retrieved.

[0178] In one embodiment, information such as task execution time, path congestion, capture success rate, capping failure, centrifugation waiting time, instrument congestion, reagent consumption trend, reasons for sample retesting, abnormal handling of hazardous materials, abnormal temperature control, and reasons for manual intervention are extracted from the historical traceability evidence chain to form a historical execution sample.

[0179] Based on historical execution samples, the execution patterns of similar tasks can be determined. For example, the grasping parameters and success rates for different types of test tubes can be statistically analyzed to adjust the initial grasping force for subsequent test tube grasping tasks; path congestion at different time periods can be statistically analyzed to adjust the robot's cross-workstation movement path; inventory warning thresholds and replenishment times can be adjusted based on reagent consumption trends; and task priorities can be adjusted based on equipment waiting conditions.

[0180] Continuous optimization is primarily used to optimize robot path planning parameters, motion execution parameters, task priorities, inventory warning thresholds, and anomaly warning thresholds, rather than directly replacing medical laboratory quality rules.

[0181] When continuous optimization results involve changes to state recognition conditions, state transition conditions, or action authorization conditions, rule adjustment suggestions should be generated first. After receiving confirmation from the quality manager or technical manager, these suggestions should then be applied to subsequent standardized operations to avoid directly changing compliance gating conditions without reviewing the learning results.

[0182] In one embodiment, the control system is connected via a communication interface to at least one of a LIS, LIMS, RIMS, instrumentation middleware, analyzer, centrifuge, smart refrigerator, and temperature control system.

[0183] Obtain specimen identification, test items, collection time, and re-examination events from LIS; obtain reagent batch number, expiration date, inventory location, and recall status from RIMS; obtain instrument availability, sample loading status, and test completion events from analyzers or instrument intermediates; and obtain the temperature and humidity status of the target work area or storage equipment from the temperature control system.

[0184] When the communication interface is unavailable, the data to be sent or received, along with its generation time, is stored in a local buffer, and the current task is controlled to execute according to preset offline rules. After the communication interface is restored, breakpoint resume transmission is performed according to the data generation order, and the local buffer state is compared with the current state of the external system. If the two are inconsistent, a state correction task or a manual review task is generated to avoid missing state changes during the communication interruption.

[0185] In one embodiment, visualized regulatory data is generated based on a digital twin, an executable task graph, and a traceable chain of evidence. This visualized regulatory data may include the current process node of the target operation object, its current compliance status, pending tasks, robot location, equipment utilization status, reagent inventory levels, hazardous materials storage status, and abnormal events.

[0186] For target work objects that are in an abnormal suspended state or awaiting manual review, the cause of the abnormality, relevant sensing data, executed actions, current location, and pending tasks can be displayed in the visualization interface, enabling authorized personnel to make processing decisions based on the complete context.

[0187] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0188] Based on the same inventive concept, this application also provides a hymenos robot control system for implementing the aforementioned hymenos robot control method for standardized medical laboratory operations. The solution provided by this system is similar to the implementation described in the above method. Therefore, the specific limitations of one or more cloud-based hymenos robot control system embodiments for standardized medical laboratory operations provided below can be found in the limitations of the hymenos robot control method for standardized medical laboratory operations described above, and will not be repeated here.

[0189] In one exemplary embodiment, such as Figure 6 As shown, this application also provides an embodied robot control system 10 for standardized operations in medical laboratories, the system comprising: The twin update module 20 is used to acquire multimodal state data of the target work object, work equipment, work space and embodied robot in the medical laboratory, and update the digital twin corresponding to the target work object, work equipment, work space and embodied robot based on the multimodal state data; The candidate action determination module 30 is used to determine the current compliance status of the target operation object based on the digital twin and the preset medical laboratory operation rules, and to determine the set of candidate actions that the embodied robot is allowed to perform based on the current compliance status and the action authorization conditions associated with the current compliance status. The task sequence generation module 40 is used to construct an executable task graph based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, as well as the candidate action set, and to generate a target task sequence based on the executable task graph. The control execution module 50 is used to control the embodied robot to perform standardized operations for the target object in different workstations according to the target task sequence, and to obtain the execution results; The evidence chain generation module 60 is used to update the digital twin and the current compliance status based on the execution result, and associate the multimodal status data, compliance status changes, action parameters and execution results to form a traceability evidence chain for standardized operations.

[0190] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0191] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0192] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps of the method described above.

[0193] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0194] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0195] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for controlling an embodied robot for standardized operations in medical laboratories, characterized in that, The method includes: Acquire multimodal state data of target work objects, work equipment, work space, and embodied robots within a medical laboratory, and update the digital twins corresponding to the target work objects, work equipment, work space, and embodied robots based on the multimodal state data; Based on the digital twin and preset medical laboratory operating rules, the current compliance status of the target operating object is determined, and based on the current compliance status and the action authorization conditions associated with the current compliance status, a set of candidate actions that the embodied robot is allowed to perform is determined. Based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, and the candidate action set, an executable task graph is constructed, and a target task sequence is generated according to the executable task graph; The robot is controlled to perform standardized tasks for the target object at different workstations according to the target task sequence, and the execution results are obtained. The digital twin and the current compliance status are updated based on the execution results, and the multimodal status data, compliance status changes, action parameters and execution results are associated to form a traceability evidence chain for the standardized operation.

2. The method according to claim 1, characterized in that, The method further includes a process for establishing the digital twin, the process of establishing the digital twin including: Data association and state parsing are performed on the multimodal state data to obtain the object state of the target operation object, the operating state of the operation equipment and the embodied robot, and the environmental state of the operation space; The object status, running status, and environment status are written into the status field of the corresponding digital twin, and the current job status of the target job object is associated with the corresponding job process node according to the business process corresponding to the target job object to obtain the digital twin.

3. The method according to claim 1, characterized in that, Based on the digital twin and preset medical laboratory operating rules, the current compliance status of the target operating object is determined, and based on the current compliance status and the action authorization conditions associated with the current compliance status, a set of candidate actions that the embodied robot is allowed to perform is determined, including: The medical laboratory operation rules are converted into status recognition conditions and action authorization conditions; The object status of the target operation object in the digital twin is matched with the status identification conditions to determine the current compliance status of the target operation object; Based on the current compliance status, query the associated action authorization conditions to obtain the allowed and prohibited actions under the current compliance status; The prohibited actions are excluded from the preset set of robot actions, and the remaining allowed actions are determined as the set of candidate actions.

4. The method according to claim 1, characterized in that, The step involves constructing an executable task graph based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, and the candidate action set, and generating a target task sequence based on the executable task graph, including: The business process corresponding to the target job object is decomposed into multiple task nodes, and task attributes are configured for each task node; wherein, the task attributes include target job object, previous state, target state, target workstation, candidate action, completion time limit and exception handling conditions; Based on the execution sequence and compliance status transition relationships between each task node, task association edges are established to obtain the initial task graph; Based on the operating status of the work equipment and the embodied robot, the set of candidate actions, and the task constraints, the executable status of each task node in the initial task graph is determined, and the executable task graph is formed. The target task sequence is extracted from the executable task graph based on the preset task priority and the execution order represented by the task association edges.

5. The method according to claim 4, characterized in that, The step of determining the executable state of each task node in the initial task graph based on the operating status of the working equipment and the android, the candidate action set, and the task constraints, to form the executable task graph, includes: For each task node in the initial task graph, determine whether the target action corresponding to the task node belongs to the candidate action set, and determine whether the operating equipment and embodied robot required to perform the target action are available based on the operating status of the operating equipment and embodied robot. When the target action belongs to the candidate action set, the required work equipment and embodied robot are available, and the task node satisfies the corresponding task constraints, the task node is determined to be in an executable state; otherwise, the task node is determined to be in a waiting state or a blocked state. The task nodes and task-related edges in the initial task graph are updated according to the executable status of each task node to form the executable task graph.

6. The method according to claim 1, characterized in that, The control of the embodied robot to perform standardized tasks for the target task object at different workstations according to the target task sequence, and to obtain the execution results, includes: Based on the target work object, target action and target workstation in the target task sequence, determine the target execution device of the embodied robot, and determine the target end effector from multiple replaceable end effectors; Based on the structure and current spatial location of the target work object, generate the initial motion parameters of the target execution device and the target end effector; The corresponding target task is executed according to the initial action parameters, and execution feedback results are obtained during the execution process; wherein, the execution feedback results include at least one of visual feedback, position feedback, force feedback and collision feedback; Adjust the initial action parameters based on the execution feedback results, and after completing the standardized operation of the current target task, jump to the next task in the target task sequence until all operations in the target task sequence are completed, and obtain the state of the target task object after execution; The execution result is determined by comparing the status of the target job object after execution with the expected completion status of the target task. The result is determined as successful execution, failed execution, or abnormal execution.

7. The method according to claim 6, characterized in that, The step of updating the digital twin and the current compliance status based on the execution result, and associating the multimodal status data, compliance status changes, action parameters, and execution results to form a traceability evidence chain for the standardized operation, includes: The actual task completion status corresponding to the target task object is determined based on the execution result. Write the actual task completion status into the digital twin; Match the actual task completion status with the status transition conditions corresponding to the current compliance status; When the state transition condition is met, the current compliance state is transitioned to the corresponding next compliance state; When the state transition conditions are not met or the execution result is execution failure or execution exception, the target job object is converted to an abnormal suspension state or a state awaiting manual review, and an exception handling task is generated. Based on the updated content, compliance status transition information, and action parameters of the digital twin, an evidence chain record corresponding to the target operation object is generated, and the association between multiple evidence chain records is established according to the execution order of each target task in the target task sequence to form a traceability evidence chain for the standardized operation.

8. A hymenoid robot control system for standardized operations in medical laboratories, characterized in that, The system includes: The twin update module is used to acquire multimodal state data of the target work object, work equipment, work space and embodied robot in the medical laboratory, and update the digital twin corresponding to the target work object, work equipment, work space and embodied robot based on the multimodal state data; The candidate action determination module is used to determine the current compliance status of the target operation object based on the digital twin and preset medical laboratory operation rules, and to determine a set of candidate actions that the embodied robot is allowed to perform based on the current compliance status and the action authorization conditions associated with the current compliance status. The task sequence generation module is used to construct an executable task graph based on the business process, task constraints, operating status of the work equipment and the embodied robot corresponding to the target work object, and the candidate action set, and to generate a target task sequence based on the executable task graph. The control execution module is used to control the embodied robot to perform standardized tasks for the target task object in different workstations according to the target task sequence, and to obtain the execution results; The evidence chain generation module is used to update the digital twin and the current compliance status based on the execution result, and associate the multimodal status data, compliance status changes, action parameters and execution results to form a traceability evidence chain for the standardized operation.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images