Body intelligent agent warehouse handling fault-tolerant control method and system and body intelligent agent

Abstract

The application provides a kind of warehousing and handling fault-tolerant control method, system and embodied intelligent agent of embodied intelligent agent, the method comprises: real-time monitoring the operating parameter of core operating component in the running process of embodied intelligent agent, wherein, core operating component is the component that influences normal operation of embodied intelligent agent;Based on operating parameter, it is judged whether there is abnormality in core operating component;In the case where there is abnormality in core operating component, based on the core operating component that exists abnormality, the abnormality level of embodied intelligent agent is determined;According to the abnormality level of embodied intelligent agent, the abnormality processing scheme matched with abnormality level is determined, and the operation of embodied intelligent agent is controlled based on abnormality processing scheme.The application realizes that the operation abnormality of embodied intelligent agent can be comprehensively monitored and independently processed, so as to improve the operation reliability and operation continuity of embodied intelligent agent in complex dynamic environment.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an embodied intelligent agent warehouse handling fault-tolerant control method, system, and embodied intelligent agent. Background Technology

[0002] With the continuous improvement of automation in logistics and warehousing, embodied intelligent agents, such as embodied robots, have been widely used in cargo handling and other operational scenarios. In actual continuous operation, due to factors such as the complex and changeable warehousing environment, the performance degradation of equipment due to long-term operation, and the diversification of cargo status, robot systems will inevitably encounter various anomalies and malfunctions.

[0003] Currently, most warehouse handling robot control systems assume that the equipment operates under "ideal conditions," lacking comprehensive real-time monitoring of the status of core components and the operational process. Anomalies are often only passively detected after hardware failures cause operational interruptions. Once an anomaly occurs, existing systems typically resort to immediate shutdown with alarms, awaiting manual intervention. This mechanism heavily relies on manual troubleshooting and recovery, resulting in long downtime, high maintenance costs, and low efficiency, failing to meet the warehousing industry's demands for continuous, large-scale operations.

[0004] Therefore, there is an urgent need for a fault-tolerant control method that can comprehensively monitor and autonomously handle anomalies in order to improve the operational reliability and work continuity of embodied robots in complex dynamic environments. Summary of the Invention

[0005] This invention provides a fault-tolerant control method, system, and embodied intelligent agent for warehouse handling, which enables comprehensive monitoring and autonomous handling of operational anomalies of the embodied intelligent agent, thereby improving the operational reliability and continuity of the embodied intelligent agent in complex dynamic environments.

[0006] This invention provides a fault-tolerant control method for embodied intelligent agents in warehouse handling. The method includes: real-time monitoring of the operating parameters of core operating components of the embodied intelligent agent during operation, wherein the core operating components are components that affect the normal operation of the embodied intelligent agent; determining whether the core operating components are abnormal based on the operating parameters; determining the abnormality level of the embodied intelligent agent based on the abnormal core operating components when the core operating components are abnormal; determining an abnormality handling scheme matching the abnormality level according to the abnormality level of the embodied intelligent agent, and controlling the operation of the embodied intelligent agent based on the abnormality handling scheme.

[0007] According to the present invention, a fault-tolerant control method for embodied intelligent agents in warehousing and handling includes determining whether the core operating components are abnormal based on the operating parameters. This includes: determining, for each core operating component, an operating parameter threshold under normal operation; determining whether the core operating component is abnormal based on the operating parameters and the operating parameter threshold, wherein the operating parameter threshold under normal operation is determined by: acquiring multiple sets of historical operating parameters of the core operating components under normal operation of the embodied intelligent agent, wherein the historical operating parameters are historical operating parameters formed by the core operating components within a preset time period, the preset time period being less than or equal to a time period threshold; and determining the operating parameter threshold under normal operation of the core operating components based on the multiple sets of historical operating parameters.

[0008] According to the present invention, a fault-tolerant control method for embodied intelligent agents in warehouse handling includes determining the anomaly level of the embodied intelligent agent based on the presence of abnormal core operating components. This includes: determining the weight value of each core operating component, wherein the weight value is positively correlated with the degree of influence of the core operating component on the normal operation of the embodied intelligent agent; assigning anomaly scores to the abnormal core operating components to obtain anomaly scores; determining the comprehensive anomaly score of the embodied intelligent agent based on the anomaly scores and the weight values ​​of the abnormal core operating components; and determining the anomaly level of the embodied intelligent agent based on the comprehensive anomaly score.

[0009] According to the present invention, a fault-tolerant control method for embodied intelligent agents in warehouse handling is provided, wherein the anomaly level includes a mild anomaly level, a moderate anomaly level, and a severe anomaly level; the step of determining the anomaly level of the embodied intelligent agent based on its comprehensive anomaly score includes: determining the anomaly level of the embodied intelligent agent as the mild anomaly level when the comprehensive anomaly score of the embodied intelligent agent is less than a first score threshold; determining the anomaly level of the embodied intelligent agent as the moderate anomaly level when the comprehensive anomaly score of the embodied intelligent agent is greater than or equal to the first score threshold and less than a second score threshold; and determining the anomaly level of the embodied intelligent agent as the severe anomaly level when the comprehensive anomaly score of the embodied intelligent agent is greater than or equal to the second score threshold.

[0010] According to a fault-tolerant control method for embodied intelligent agents in warehouse handling provided by the present invention, the step of determining an anomaly handling scheme matching the anomaly level of the embodied intelligent agent includes: in response to the anomaly level of the embodied intelligent agent being a mild anomaly level, determining the anomaly handling scheme as a silent repair handling scheme, wherein the silent repair handling scheme is used to characterize the calibration of the core operating components with an anomaly without interrupting the operation of the embodied intelligent agent; in response to the anomaly level of the embodied intelligent agent being a moderate anomaly level, determining the anomaly handling scheme as an active intervention handling scheme, wherein the active intervention handling scheme is used to characterize the calibration of the core operating components with an anomaly while interrupting the operation of the embodied intelligent agent; and in response to the anomaly level of the embodied intelligent agent being a severe anomaly level, determining the anomaly handling scheme as a redundancy switching handling scheme, wherein the redundancy switching handling scheme is used to characterize the calling of a backup embodied intelligent agent with the same operational capabilities as the embodied intelligent agent.

[0011] According to the present invention, a fault-tolerant control method for embodied intelligent agents in warehousing and handling is provided. The method for calibrating the malfunctioning core operating component without interrupting the operation of the embodied intelligent agent is implemented in the following manner: without interrupting the operation of the embodied intelligent agent, a reference operating parameter for the malfunctioning core operating component is determined; based on the reference operating parameter, the malfunctioning core operating component is calibrated. The reference operating parameter is determined by: acquiring multiple sets of historical operating parameters of the core operating component when the embodied intelligent agent performs the same task and operates normally, and determining the reference operating parameter based on the historical operating parameters; or acquiring multiple sets of experimental operating parameters of the core operating component when the embodied intelligent agent performs the same task and operates normally in an experimental environment, and determining the reference operating parameter based on the experimental operating parameters.

[0012] According to the present invention, a fault-tolerant control method for embodied intelligent agents in warehousing and handling includes calibrating the core operating components that are abnormal when the operation of the embodied intelligent agent is interrupted. This includes calling the core operating components to repeatedly execute operating actions multiple times within a preset number of times until the operating actions meet preset requirements, so as to calibrate the core operating components that are abnormal.

[0013] According to the present invention, a fault-tolerant control method for embodied intelligent agents in warehousing and handling is provided. The method further includes: if the operation action obtained after repeatedly executing the operation action of the core operating component within a preset number of times does not meet the preset requirements, calling a backup embodied intelligent agent with the same operation capability as the embodied intelligent agent to operate.

[0014] This invention also provides an embodied intelligent agent warehouse handling fault-tolerant control system, the system comprising: a monitoring module for real-time monitoring of the operating parameters of core operating components during the operation of the embodied intelligent agent, wherein the core operating components are components that affect the normal operation of the embodied intelligent agent; a judgment module for determining whether the core operating components are abnormal based on the operating parameters; a determination module for determining the abnormality level of the embodied intelligent agent based on the abnormal core operating components when the core operating components are abnormal; and a processing module for determining an abnormality handling scheme matching the abnormality level of the embodied intelligent agent, and controlling the operation of the embodied intelligent agent based on the abnormality handling scheme.

[0015] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the embodied intelligent agent storage and handling fault-tolerant control method as described above.

[0016] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the embodied intelligent agent warehousing and handling fault-tolerant control method as described above.

[0017] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the embodied intelligent agent storage and handling fault-tolerant control method as described above.

[0018] This invention provides a fault-tolerant control method, system, and embodied intelligent agent for warehouse handling. The method includes: real-time monitoring of the operating parameters of core operating components of the embodied intelligent agent during operation, wherein the core operating components are those affecting the normal operation of the embodied intelligent agent; determining whether there is an anomaly in the core operating components based on the operating parameters; determining the anomaly level of the embodied intelligent agent based on the anomaly in the core operating components; determining an anomaly handling scheme matching the anomaly level, and controlling the operation of the embodied intelligent agent based on the anomaly handling scheme. This invention enables comprehensive monitoring and autonomous handling of operational anomalies of the embodied intelligent agent, thereby improving the operational reliability and operational continuity of the embodied intelligent agent in complex dynamic environments. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0020] Figure 1 This is a flowchart illustrating the embodied intelligent agent warehousing and handling fault-tolerant control method provided by the present invention.

[0021] Figure 2 This is a flowchart illustrating the process of determining the anomaly level of the embodied intelligent agent based on the presence of an anomaly in the core operating components provided by the present invention.

[0022] Figure 3 This is a schematic diagram of the embodied intelligent agent warehouse handling fault-tolerant control system provided by the present invention.

[0023] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0025] This invention provides an embodied intelligent agent-based fault-tolerant control method for warehouse handling. By constructing a multi-dimensional, end-to-end anomaly monitoring system, it scientifically classifies anomalies and faults, matches them with differentiated adaptive processing strategies, and combines task self-recovery and self-learning optimization mechanisms. This minimizes operational interruptions, reduces human intervention, and improves the overall reliability and robustness of the robot system. This invention can be applied to control systems employing the embodied intelligent agent-based fault-tolerant control method for warehouse handling.

[0026] Figure 1 This is a flowchart illustrating the embodied intelligent agent warehousing and handling fault-tolerant control method provided by the present invention.

[0027] The following will combine Figure 1 The process of the embodied intelligent agent warehouse handling fault-tolerant control method provided by the present invention will be described.

[0028] In an exemplary embodiment of the present invention, combined with Figure 1As can be seen, the embodied intelligent agent warehouse handling fault-tolerant control method may include steps 110 to 140, and each step will be described below.

[0029] In step 110, the operating parameters of the core operating components of the embodied intelligent agent are monitored in real time during operation. The core operating components are those that affect the normal operation of the embodied intelligent agent.

[0030] In one embodiment, the embodied intelligent agent can be a warehouse handling robot equipped with an autonomous mobile chassis, a robotic arm, grippers, and multimodal sensors. The core operating components are those that affect the normal operation of the embodied intelligent agent. These core operating components may include vision sensors (RGB-D cameras), LiDAR, force sensors (six-dimensional force sensors), joint motors, chassis drive motors, etc. During application, the operating parameters of the core operating components can be monitored in real time during the operation of the embodied intelligent agent.

[0031] For vision sensors, the operating parameters can be image sharpness score (based on gradient amplitude) and rate of change between consecutive frames; for lidar, the operating parameters can be point cloud echo intensity and noise point ratio; for force sensors, the operating parameters can be zero bias and end load force fluctuation amplitude; for articulated motors, the operating parameters can be current, temperature, and encoder position error; for chassis motors, the operating parameters can be code disk pulse count and left and right wheel speed difference.

[0032] In step 120, based on the operating parameters, it is determined whether there is any abnormality in the core operating components.

[0033] In one embodiment, a dynamic threshold (such as ±3 times the standard deviation of the historical mean) can be set for each operating parameter, and anomaly detection can be performed in conjunction with the 3σ principle.

[0034] In step 130, if there is an anomaly in the core operating component, the anomaly level of the embodied intelligent agent is determined based on the anomaly in the core operating component.

[0035] In step 140, an anomaly handling scheme matching the anomaly level is determined based on the anomaly level of the embodied agent, and the operation of the embodied agent is controlled based on the anomaly handling scheme.

[0036] In one embodiment, when a core operating component malfunctions, the malfunction level of the embodied agent can be determined based on the malfunctioning core operating component. Further, based on the embodied agent's malfunction level, a matching malfunction handling scheme is determined, and the embodied agent's operation is controlled according to the malfunction handling scheme. After executing the malfunction handling scheme, if the malfunction has been resolved but the task has not been completed, the system automatically resumes the operation from the interruption point, achieving seamless transition.

[0037] In this embodiment, by monitoring the operating parameters of core components in real time and detecting early signs of anomalies, combined with a hierarchical processing mechanism, anomalies can be autonomously repaired without interruption or with only a short pause, avoiding the shutdown of the entire production line, thereby supporting 24 / 7 continuous operation in the warehousing scenario.

[0038] This invention provides a fault-tolerant control method for embodied intelligent agents in warehouse handling. The method includes: real-time monitoring of the operating parameters of core operating components during the operation of the embodied intelligent agent, wherein the core operating components are those affecting the normal operation of the embodied intelligent agent; determining whether there is an anomaly in the core operating components based on the operating parameters; determining the anomaly level of the embodied intelligent agent based on the anomaly in the core operating components; determining an anomaly handling scheme matching the anomaly level, and controlling the operation of the embodied intelligent agent based on the anomaly handling scheme. This invention enables comprehensive monitoring and autonomous handling of operational anomalies of embodied intelligent agents, thereby improving the operational reliability and operational continuity of embodied intelligent agents in complex dynamic environments.

[0039] In yet another exemplary embodiment of the present invention, the following description is used to illustrate the determination of whether there is an abnormality in the core operating component based on the operating parameters: For each core operating component, determine the threshold values ​​of its operating parameters under normal operating conditions; Based on operating parameters and operating parameter thresholds, it is determined whether there are any abnormalities in the core operating components. The operating parameter thresholds of the core operating components under normal operation are determined in the following way: The system acquires multiple sets of historical operating parameters of the core operating components of the embodied intelligent agent under normal operating conditions. The historical operating parameters are the historical operating parameters formed by the core operating components within a preset time period, where the preset time period is less than or equal to the time threshold. Based on multiple sets of historical operating parameters, the operating parameter thresholds of core operating components under normal operating conditions are determined.

[0040] In one embodiment, taking the "chassis drive motor" in an embodied intelligent agent as an example, its operating parameter during normal operation is the "motor current". The following steps can be used to dynamically calculate the threshold of the operating parameter of this component.

[0041] During the period when the robot completes a standard handling task without any fault records (e.g., the stable operation window of the last 30 seconds, with a duration threshold set to 60 seconds), the current value of the chassis drive motor can be continuously collected at a sampling rate of 100Hz, resulting in a total of 3000 sets of historical operating parameters.

[0042] Statistical analysis can be performed on the aforementioned 3000 sets of current data to calculate their mean μ and standard deviation σ. Subsequently, the 3σ principle is used to set the threshold values ​​for normal operating parameters: lower threshold = μ - 3σ; upper threshold = μ + 3σ. Based on this, the operating parameter threshold values ​​for core operating components under normal operating conditions can be determined.

[0043] In another embodiment, during subsequent continuous operation, the current value of the chassis drive motor (real-time operating parameter) is collected in real time and compared with the aforementioned threshold range. If the current value is within the range of the operating parameter threshold, it proves that the core operating component (motor) is normal; if the current value is not within the range of the operating parameter threshold, it proves that the core operating component (motor) is abnormal.

[0044] To further improve the adaptability of the threshold, the historical dataset can be updated every "pre-set time interval" (e.g., every 30 seconds), and μ and σ can be recalculated to dynamically adjust the operating parameter thresholds to adapt to the performance degradation of the motor due to long-term operation or changes in ambient temperature. This update mechanism ensures the real-time performance and accuracy of the threshold.

[0045] In this embodiment, the threshold is dynamically calculated based on the embodied intelligent agent's recent historical operating parameters, fully considering the individual differences of the robot and its real-time operating conditions. This makes anomaly detection more accurate and significantly reduces false alarm and false negative rates. By limiting the historical operating parameters to "previously preset duration less than or equal to a duration threshold," the threshold calculation relies solely on recent normal operating data, which can quickly reflect slow drifts in system state (such as motor aging or sensor zero bias changes), while avoiding the use of outdated data that leads to sluggish threshold response. The upper bound constraint on the duration threshold prevents the data window from becoming too long and losing its timeliness.

[0046] Figure 2 This is a flowchart illustrating the process of determining the anomaly level of the embodied intelligent agent based on the presence of an anomaly in the core operating components provided by the present invention.

[0047] The following will combine Figure 2 The process provided by this invention for determining the anomaly level of the embodied intelligent agent based on the presence of anomalies in the core operating components is described.

[0048] In an exemplary embodiment of the present invention, combined with Figure 2 As can be seen, determining the abnormality level of the embodied intelligent agent based on the existence of abnormal core operating components may include steps 210 to 240, and each step will be described below.

[0049] In step 210, the weight values ​​of each core operating component are determined, wherein the weight values ​​are positively correlated with the degree of influence of the core operating component on the normal operation of the embodied intelligent agent.

[0050] In one embodiment, for an embodied intelligent agent (warehouse handling robot), the degree of influence of each component on the normal operation of the robot can be preset, and weight values ​​can be assigned accordingly. The weight value is positively correlated with the degree of influence, that is, the more critical the component, the higher the weight. In one example, the weight values ​​can be pre-calibrated using the Analytic Hierarchy Process (AHP) combined with expert experience and stored in the robot's edge computing unit.

[0051] In step 220, an anomaly score is assigned to the core operating components that have an anomaly, thus obtaining the anomaly score of the core operating components that have an anomaly.

[0052] In one embodiment, an exception score can be assigned to a core operating component that exhibits an anomaly, thus obtaining the exception score of the core operating component with the anomaly. In one example, for a core operating component determined to have an anomaly, its exception score can be directly assigned the value 1; for a core operating component determined not to have an anomaly, its exception score can be directly assigned the value 0.

[0053] In another example, the anomaly score can be determined based on the duration and magnitude of the anomaly of the core operating component, where the longer the duration of the anomaly and / or the greater the magnitude of the deviation, the higher the anomaly score.

[0054] In step 230, the overall anomaly score of the embodied agent is determined based on the anomaly score of the core operating components that have anomalies and the weight value of the core operating components that have anomalies.

[0055] In step 240, the abnormality level of the embodied agent is determined based on the overall abnormality score of the embodied agent.

[0056] In another embodiment, the overall anomaly score of the embodied agent can be determined based on the anomaly scores of the core operating components that are abnormal and the weighted sum of the weight values ​​of the core operating components that are abnormal.

[0057] Furthermore, based on the obtained comprehensive abnormality score of the embodied intelligent agent, the abnormality level of the embodied intelligent agent is classified.

[0058] In another exemplary embodiment of the present invention, continuing with the previously described embodiments, the anomaly level may include a mild anomaly level, a moderate anomaly level, and a severe anomaly level; wherein, determining the anomaly level of the embodied intelligent agent based on its comprehensive anomaly score can be achieved in the following manner: If the overall anomaly score of the embodied agent is less than the first score threshold, the anomaly level of the embodied agent is determined to be mild anomaly level. If the overall anomaly score of the embodied agent is greater than or equal to the first score threshold and less than the second score threshold, the anomaly level of the embodied agent is determined to be moderate anomaly level. If the overall anomaly score of the embodied agent is greater than or equal to the second score threshold, the anomaly level of the embodied agent is determined to be severe anomaly level.

[0059] In one embodiment, a first score threshold and a second score threshold can be preset based on historical fault data and expert experience, wherein the first score threshold is lower than the second score threshold. The first and second score thresholds can be dynamically adjusted according to different warehousing scenarios, robot models, or operational safety requirements.

[0060] In the application process, if the overall abnormality score of the embodied agent is less than the first score threshold, the abnormality level of the embodied agent can be determined as mild abnormality; if the overall abnormality score of the embodied agent is greater than or equal to the first score threshold and less than the second score threshold, the abnormality level of the embodied agent can be determined as moderate abnormality; if the overall abnormality score of the embodied agent is greater than or equal to the second score threshold, the abnormality level of the embodied agent can be determined as severe abnormality.

[0061] In this embodiment, by setting a first score threshold and a second score threshold, the continuous comprehensive anomaly score is discretized into three distinct anomaly levels (mild, moderate, and severe). Each level corresponds to a differentiated fault-tolerant handling scheme, avoiding a "one-size-fits-all" shutdown and achieving optimal resource allocation: mild problems are silently repaired without interrupting operations, moderate problems are actively intervened with a short pause, and severe problems are safely shut down and redundantly switched over.

[0062] In an exemplary embodiment of the present invention, determining an anomaly handling scheme matching the anomaly level of the embodied intelligent agent can be achieved in the following manner: In response to the fact that the anomaly level of the embodied intelligent agent is a mild anomaly level, the anomaly handling scheme is determined to be a silent repair scheme. The silent repair scheme is used to characterize the calibration of the core operating components with anomalies without interrupting the operation of the embodied intelligent agent. In response to the embodied intelligent agent's anomaly level being moderate, the anomaly handling scheme is determined to be an active intervention scheme. The active intervention scheme is used to characterize the calibration of the core operating components with anomalies when the operation of the embodied intelligent agent is interrupted. In response to the abnormality level of the embodied intelligent agent being classified as severe abnormality, the abnormality handling scheme is determined to be a redundancy switching scheme. The redundancy switching scheme is used to indicate that a backup embodied intelligent agent with the same operational capabilities as the embodied intelligent agent is invoked for operation.

[0063] In one embodiment, when the anomaly level of the embodied agent is mild, the anomaly handling scheme can be determined to be a silent repair scheme. With the silent repair scheme, the core operating components exhibiting the anomaly can be calibrated without interrupting the operation of the embodied agent.

[0064] In one example, while transporting goods, the embodied intelligent agent (robot A) experiences a sudden surge in the instantaneous current of its chassis drive motor from the normal value of 2.5A to 3.3A (a slight overcurrent). The overall anomaly score is calculated to be 20 points, which is less than the first score threshold of 25, thus classifying it as a minor anomaly. Given that the embodied intelligent agent's anomaly level is minor, the appropriate anomaly handling solution is a silent repair solution.

[0065] For the silent repair solution, the system does not pause any of the robot A's actions (maintaining the moving speed and the robotic arm's posture); in the motion control loop, the current signal can be smoothed and filtered by a Kalman filter to remove spike interference; at the same time, the PID controller dynamically adjusts the PWM duty cycle of the drive motor according to the current load to bring the current back to the normal range; the entire calibration process is completed in the background, and the robot A continues to transport goods without the operator or the host computer noticing anything, resulting in zero interruption of the operation.

[0066] The essence of the silent repair solution lies in eliminating the impact of abnormalities by performing online calibration of abnormal components through software algorithms (filtering, parameter fine-tuning) without interrupting operation.

[0067] In another embodiment, when the anomaly level of the embodied agent is moderate, the anomaly handling scheme can be determined as an active intervention scheme. For the active intervention scheme, the core operating components exhibiting the anomaly can be calibrated while the operation of the embodied agent is interrupted.

[0068] In one example, when robot A grasps goods, the confidence level of the visual sensor drops to 55% (moderate anomaly), and the force sensor reports insufficient grasping force (slight anomaly). The overall anomaly score is calculated to be 38 points, falling within the range [25, 60), and is thus classified as a moderate anomaly. When the anomaly level of the embodied agent is moderate, the anomaly handling plan can be determined to be an active intervention plan.

[0069] For proactive intervention solutions, the system can immediately interrupt the current operation of robot A: pause the closing action of the robotic arm gripper and stop the chassis movement; it can invoke the subtask retry logic, specifically including: controlling robot A to move slightly backward and readjusting the perspective to obtain a clearer image of the goods; recalculating the coordinates of the gripping point based on the new visual feedback; adjusting the gripper opening width and the target value of the gripping force, and re-executing the gripping action; after calibration, the system reads the breakpoint status from the task stack (e.g., "grabbing goods"), automatically resumes operation, and continues the handling task.

[0070] The essence of the proactive intervention solution is to briefly interrupt operation and calibrate abnormal components through retry, parameter adjustment, or local path replanning. The task will automatically resume after successful handling.

[0071] In another embodiment, when the anomaly level of the embodied agent is a severe anomaly level, the anomaly handling scheme can be determined to be a redundancy switching scheme. For the redundancy switching scheme, a backup embodied agent with the same operational capabilities as the embodied agent can be called upon to run.

[0072] In one example, the joint motors of robot A's robotic arm suddenly jammed (completely failed), and the vision sensor output went black. The overall anomaly score was calculated to be 82, which is greater than or equal to the second score threshold of 60, thus classifying it as a severe anomaly. Given that the embodied agent's anomaly level is severe, the anomaly handling scheme can be determined to be a redundant switching scheme.

[0073] For redundancy switching, the system immediately triggers safety protection mechanisms, such as robotic arm brakes and chassis emergency stops, to prevent the accident from escalating. Additionally, the current task context of robot A (cargo ID, gripping posture, target location, traveled path, cargo weight, etc.) is serialized and encrypted with a signature; this is then sent to the central dispatch system via wireless network. After parsing and verification, the central dispatch system retrieves a backup embodied intelligent agent (robot B) with the same operational capabilities as robot A from the idle queue. Robot B navigates to the location of robot A, confirms its identity by scanning the cargo's QR code, uses visual repositioning to confirm the gripping point, safely retrieves the cargo, and continues transporting it to the target location. In another example, robot A's fault logs can be reported to the maintenance platform, notifying manual repair personnel.

[0074] The essence of the redundancy switching solution is that when the current embodied intelligent agent exits operation, another standby intelligent agent with the same operational capabilities is called to take over the task, thereby achieving cluster-level redundancy and fault tolerance.

[0075] In this embodiment, minor anomalies are handled with silent repair, ensuring continuous operation 24 / 7 without interruption; moderate anomalies are handled with proactive intervention, resulting in only brief pauses to prevent small problems from escalating into serious failures; and severe anomalies are handled with redundant switching, sacrificing individual machines but ensuring uninterrupted operation of the overall warehousing system. These three approaches each fulfill their specific roles, avoiding the unreasonable resource allocation issues of "treating small problems with major solutions" or "treating major problems with minor solutions."

[0076] In yet another exemplary embodiment of the present invention, calibrating core operating components that exhibit abnormalities without interrupting the operation of the embodied intelligent agent can be achieved in the following manner: Without interrupting the operation of the embodied intelligent agent, determine the reference operating parameters of the core operating components where anomalies exist; Based on reference operating parameters, the core operating components exhibiting abnormalities are calibrated. These reference operating parameters are determined using the following method: Obtain multiple sets of historical operating parameters of the core operating components of the embodied intelligent agent when performing the same task and operating normally, and determine reference operating parameters based on the historical operating parameters, or Multiple sets of experimental operating parameters of the core operating components are obtained when the embodied intelligent agent performs the same task and operates normally in the experimental environment, and reference operating parameters are determined based on the experimental operating parameters.

[0077] In one embodiment, without interrupting the operation of the embodied intelligent agent, for example, without pausing the robot's movement or the robotic arm's actions, and while maintaining operational continuity, reference operating parameters of the core operating components where anomalies occur are determined.

[0078] In one embodiment, the embodied agent has accumulated a certain amount of fault-free operational data in the same warehouse handling tasks. During application, the locally stored historical operation database can be retrieved to filter out time windows that are of the same task type as the current task (e.g., "moving type X goods from shelf A to location B") and are determined to be operating normally (without any abnormal records). Within this window, multiple sets of historical operational parameters of the core operating components (e.g., image sharpness scores of the vision sensor) are extracted. Example: In the last 100 successful handling tasks, the sharpness score of the vision sensor remained stable between 92 and 97, with a sampling interval of 0.1 seconds, resulting in 100,000 sets of data. Statistical characteristic values ​​(such as median or mean) of these historical operational parameters can be calculated as reference operational parameters. For example, the calculated reference sharpness is 94.5.

[0079] In another embodiment, the embodied agent can be placed in a standard experimental environment similar to that of a real warehouse setting in terms of temperature, lighting, and ground friction coefficient. The robot is controlled to perform a standardized task identical to the task to be calibrated (e.g., moving at a fixed speed, grasping standard goods). During the experiment, multiple sets of experimental operating parameters are recorded for the core operating components under normal operating conditions. For example, 20 consecutive handling tasks are performed in the experimental environment, and the sharpness scores of the visual sensors are collected, resulting in 2000 sets of data. Statistical characteristic values ​​(such as the mean or truncated mean) of these experimental operating parameters are calculated as reference operating parameters. For example, the reference sharpness is 95.0.

[0080] In the application, taking the example of a "decrease in visual sensor image clarity" (mild anomaly) during material handling, the system can obtain a reference clarity of 94.5 from historical operating parameters. The system does not pause the robot's movement or the robotic arm's actions, maintaining operational continuity. In the background, a silent repair algorithm is automatically invoked: based on the difference between the reference clarity (94.5) and the current clarity (e.g., 78), the camera's automatic gain control (AGC) and exposure time parameters are adjusted to gradually restore image quality.

[0081] In addition, for persistent, brief, or occasional loss of clarity (such as dust obscuring the image), the system can use sliding window filtering, temporarily using reference operating parameters as ideal values ​​to enhance the current image, and automatically resume normal acquisition after the anomaly disappears.

[0082] Through the aforementioned process, the core operating components with abnormalities were calibrated without interruption of the operation and without the operator's awareness.

[0083] In another embodiment, after calibration, the system continues to monitor the operating parameters of the component in real time. If it deviates from the reference operating parameters again, the silent repair process described above is repeated; if the degree of deviation worsens or the frequency is too high, a level upgrade (moderate anomaly) can be triggered, at which point an active intervention plan will be invoked.

[0084] In existing technologies, even minor anomalies such as sensor drift or image blurring often require pausing the robot for recalibration or manual cleaning. This embodiment, by acquiring reference operating parameters, completes calibration without stopping operation, ensuring the continuity of handling tasks, and is particularly suitable for high-frequency warehousing scenarios operating 24 / 7.

[0085] In yet another exemplary embodiment of the present invention, when the operation of the embodied intelligent agent is interrupted, the calibration of the core operating components that are malfunctioning can be achieved in the following manner: In the event of an interruption in the operation of the embodied intelligent agent, the core operating components are invoked to repeatedly execute the operating actions within a preset number of times until the operating actions meet the preset requirements, so as to calibrate the core operating components that are abnormal.

[0086] In one embodiment, taking a moderate anomaly encountered by an embodied intelligent agent (warehouse handling robot) while performing a "grabbing goods" task as an example, the system can trigger an active intervention scheme to immediately interrupt the robot's current operation (pause the gripper action, brake the chassis, and maintain the position of the robotic arm). Furthermore, the system can call upon core operating components to repeatedly execute the operation action multiple times within a preset number of times until the operation action meets preset requirements. For the "gripper grasping" operation action, the preset requirements can be successful grasping and stable goods, with the judgment criteria being that the force sensor feedback grasping force reaches more than 90% of the target value, and visual servo confirms that the goods are not slipping and the posture deviation is less than 2 degrees.

[0087] The preset number of attempts can be the maximum number of retries, for example, it can be 3 times.

[0088] In one possible scenario, for the first repetition, the system controls the robotic arm to fully release the gripper and return to the initial gripping position; adjust the gripping strategy, which can be done by adjusting the gripper's rotation joints according to the visual feedback of the cargo's deflection angle, aligning the gripper with the long side of the cargo; and then re-execute the gripper closing action. If the operation is determined not to meet the preset requirements, a second repetition can be performed. For example, it can be interrupted and returned again; the strategy can be adjusted, in addition to posture alignment, by reducing the gripper closing speed by 30% to make force control smoother; at the same time, the gripper opening width can be reduced by 2mm to increase the contact surface pressure; and then the closing action can be re-executed. If the operation is determined to meet the preset requirements, the system stops retrying and does not attempt a third retry.

[0089] After calibration, the system reads the breakpoint from the task stack ("Goods grabbed, ready to be moved") and automatically resumes robot operation: the robotic arm lifts the goods, and the chassis begins to move towards the target storage location. The entire process involves very short interruptions and requires no manual intervention.

[0090] It should be noted that the operating actions are not limited to grasping actions, but can also be applied to other components that require repeated calibration. For a mobile chassis as the core operating component with an anomaly, the operating action can be straight-line driving; the preset requirement is a path tracking error of <3cm; the repetitive execution method can be repeatedly executing "forward 2 meters - stop - backtrack - replan". For a vision servo system as the core operating component with an anomaly, the operating action can be target recognition and positioning; the preset requirement is a recognition confidence level >85%; the repetitive execution method can be repeatedly executing "small movement - reacquire image - recognition". For a robotic arm joint as the core operating component with an anomaly, the operating action can be point-to-point movement; the preset requirement is an end-effector positioning error of <2mm; the repetitive execution method can be repeatedly executing "return to zero - move to target point" and compensating successively.

[0091] In traditional methods, when a moderate anomaly occurs (such as grasping failure or positioning deviation), the robot is usually stopped and an alarm is triggered, waiting for manual repositioning of the goods or adjustment of the robot's posture. This embodiment, by allowing the robot to automatically repeat its actions and adaptively adjust its strategy (such as changing posture, speed, and force) in each attempt, can complete calibration without human intervention, significantly shortening recovery time.

[0092] In yet another exemplary embodiment of the present invention, continuing with the previously described embodiments, the method further includes: If the execution actions obtained after repeatedly executing the core operating components within a preset number of times do not meet the preset requirements, a backup embodied intelligent agent with the same operational capabilities as the embodied intelligent agent will be invoked to perform the operation.

[0093] In one embodiment, continuing with the example described above, if the preset requirements cannot be met after reaching the preset limit of the number of times (e.g., 3 times), the system will consider that the moderate anomaly cannot be repaired by active intervention. At this time, the anomaly level will be automatically upgraded to a severe anomaly, and a redundancy switching processing scheme will be triggered (calling a backup robot).

[0094] During application, Robot A completely stops all movements, engaging the robotic arm brakes and chassis brakes to prevent goods from falling or equipment damage. It also serializes key task information, including the goods' unique identifier, current location, target location, records of attempted grasping strategies, physical attributes such as weight and dimensions, and task breakpoint status (e.g., "Grab unsuccessful, goods still on the shelf"). The packaged data is then encrypted, signed, and sent to the central dispatch system.

[0095] The central scheduling system can locate a robot (Robot B) with the same operational capabilities as Robot A in the idle queue. Robot B possesses the same type of robotic arm, gripper, and force control precision. The scheduling system then sends the task context to Robot B. Robot B navigates to the current location of the goods and confirms its identity by scanning the goods' unique identifier. Using visual relocation and force guidance, it re-executes the grasping action (since the goods have not been moved, the grasping conditions are the same as Robot A's first attempt, but Robot B succeeds using default parameters). After successful grasping, Robot B transports the goods to the target location along the original task route.

[0096] In another embodiment, the detailed fault log of robot A (including the parameters of the three repeated executions and the reasons for failure) can also be uploaded to the operation and maintenance platform to notify the human to perform subsequent maintenance on robot A.

[0097] In this embodiment, the redundancy switchover is triggered after a moderate anomaly fails to be addressed through proactive intervention (repeated execution), rather than being initially classified as a severe anomaly. This tiered escalation mechanism avoids prematurely abandoning repairable robots while ensuring timely switching when proactive intervention proves ineffective, thus guaranteeing operational continuity.

[0098] As described above, the embodied intelligent agent-based fault-tolerant control method for warehouse handling provides anomalies that are graded and resolved in a manner that requires no downtime or only brief pauses by handling most level 1 and 2 anomalies (corresponding to mild and moderate anomaly levels, respectively). Combined with a task self-recovery mechanism, this reduces the overall interruption rate of handling operations, truly achieving 24 / 7 continuous operation. It also autonomously handles the vast majority of common anomalies, requiring manual intervention only in cases of level 3 severe faults. The fault self-recovery function also avoids the tedious steps of manual recovery, significantly reducing the input and labor intensity of on-site maintenance personnel. A comprehensive monitoring mechanism ensures timely detection of anomalies and faults; differentiated handling strategies ensure efficient resource utilization, avoiding "over-treatment" of minor problems and "under-treatment" of major problems; and a self-learning mechanism allows the system to continuously improve from experience, resulting in sustained improvement in system stability and reliability over long-term operation, effectively reducing the risk of cargo damage and secondary equipment failures. When a single robot experiences a Level 3 fault (corresponding to a severe anomaly), the task can be seamlessly switched to other healthy robots through task context serialization and safe handover, demonstrating cluster-level fault tolerance and ensuring uninterrupted operation of the entire warehousing system.

[0099] The embodied intelligent agent storage and handling fault-tolerant control system provided by the present invention is described below. The embodied intelligent agent storage and handling fault-tolerant control system described below can be referred to in correspondence with the embodied intelligent agent storage and handling fault-tolerant control method described above.

[0100] Figure 3 This is a schematic diagram of the embodied intelligent agent warehouse handling fault-tolerant control system provided by the present invention.

[0101] The following will combine Figure 3 The structure of the embodied intelligent agent warehouse handling fault-tolerant control system provided by the present invention will be described.

[0102] In yet another exemplary embodiment of the present invention, combined with Figure 3 As can be seen, the embodied intelligent agent warehouse handling fault-tolerant control system may include a monitoring module 310, a judgment module 320, a determination module 330, and a processing module 340. Each module will be introduced below.

[0103] The monitoring module 310 can be configured to monitor in real time the operating parameters of the core operating components of the embodied intelligent agent during operation, wherein the core operating components are the components that affect the normal operation of the embodied intelligent agent; The judgment module 320 can be configured to determine whether there is an abnormality in the core operating component based on the operating parameters; The determination module 330 can be configured to determine the anomaly level of the embodied intelligent agent based on the anomaly of the core operating component when the core operating component is abnormal. The processing module 340 can be configured to determine an anomaly handling scheme that matches the anomaly level of the embodied intelligent agent, and control the operation of the embodied intelligent agent based on the anomaly handling scheme.

[0104] In an exemplary embodiment of the present invention, the determination module 320 can determine whether the core operating component is abnormal based on the operating parameters in the following manner: For each of the core operating components, determine the operating parameter thresholds of the core operating components under normal operating conditions; Based on the operating parameters and the operating parameter thresholds, it is determined whether the core operating component is abnormal. The operating parameter thresholds for the core operating component under normal operating conditions are determined using the following method: The system acquires multiple sets of historical operating parameters of the core operating components under normal operating conditions of the embodied intelligent agent, wherein the historical operating parameters are the historical operating parameters formed by the core operating components under a preset time period, and the preset time period is less than or equal to a time threshold. Based on multiple sets of historical operating parameters, the operating parameter thresholds of the core operating components under normal operating conditions are determined.

[0105] In an exemplary embodiment of the present invention, the determining module 330 may determine the anomaly level of the embodied intelligent agent based on the presence of an anomaly in the core operating component in the following manner: Determine the weight value of each of the core operating components, wherein the weight value is positively correlated with the degree of influence of the core operating component on the normal operation of the embodied intelligent agent; Assign an anomaly score to the core operating component that has an anomaly, and obtain the anomaly score of the core operating component that has an anomaly; Based on the anomaly scores of the core operating components that are abnormal, and the weight values ​​of the core operating components that are abnormal, the overall anomaly score of the embodied intelligent agent is determined. The abnormality level of the embodied agent is determined based on its comprehensive abnormality score.

[0106] In an exemplary embodiment of the present invention, the anomaly level includes a mild anomaly level, a moderate anomaly level, and a severe anomaly level; the determining module 330 can determine the anomaly level of the embodied intelligent agent based on the comprehensive anomaly score of the embodied intelligent agent in the following manner: If the overall anomaly score of the embodied agent is less than the first score threshold, the anomaly level of the embodied agent is determined to be the mild anomaly level. If the overall anomaly score of the embodied agent is greater than or equal to the first score threshold and less than the second score threshold, the anomaly level of the embodied agent is determined to be the moderate anomaly level. If the overall anomaly score of the embodied agent is greater than or equal to the second score threshold, the anomaly level of the embodied agent is determined to be the severe anomaly level.

[0107] In an exemplary embodiment of the present invention, the processing module 340 may determine an anomaly handling scheme matching the anomaly level of the embodied intelligent agent in the following manner: In response to the abnormality level of the embodied intelligent agent being the mild abnormality level, the abnormality handling scheme is determined to be a silent repair handling scheme, wherein the silent repair handling scheme is used to characterize the calibration of the core operating components with abnormality without interrupting the operation of the embodied intelligent agent; In response to the abnormality level of the embodied intelligent agent being the moderate abnormality level, the abnormality handling scheme is determined to be an active intervention handling scheme, wherein the active intervention handling scheme is used to characterize the calibration of the core operating components with abnormality when the operation of the embodied intelligent agent is interrupted; In response to the abnormality level of the embodied intelligent agent being the severe abnormality level, the abnormality handling scheme is determined to be a redundancy switching scheme, wherein the redundancy switching scheme is used to indicate that a backup embodied intelligent agent with the same operational capabilities as the embodied intelligent agent is invoked for operation.

[0108] In an exemplary embodiment of the present invention, the processing module 340 can perform calibration on the abnormal core operating components without interrupting the operation of the embodied intelligent agent in the following manner: Without interrupting the operation of the embodied intelligent agent, the reference operating parameters of the core operating components that are malfunctioning are determined; Based on the reference operating parameters, the core operating components exhibiting abnormalities are calibrated, wherein the reference operating parameters are determined in the following manner: Obtain multiple sets of historical operating parameters of the core operating components of the embodied intelligent agent when performing the same task and operating normally, and determine the reference operating parameters based on the historical operating parameters, or Multiple sets of experimental operating parameters of the core operating components of the embodied intelligent agent are obtained when the agent performs the same task in the experimental environment and operates normally. Based on the experimental operating parameters, the reference operating parameters are determined.

[0109] In an exemplary embodiment of the present invention, the processing module 340 may perform calibration on the abnormal core operating components in the event that the operation of the embodied intelligent agent is interrupted, in the following manner: In the event that the operation of the embodied intelligent agent is interrupted, the core operating component is invoked to repeatedly execute the operating action within a preset number of times until the operating action meets the preset requirements, so as to calibrate the core operating component that has an abnormality.

[0110] In an exemplary embodiment of the present invention, the processing module 340 may further be configured to: If the operation results obtained after repeatedly executing the operation action by the core operating component within a preset number of times do not meet the preset requirements, a backup embodied intelligent agent with the same operational capabilities as the embodied intelligent agent is invoked to perform the operation.

[0111] Based on the same inventive concept, this invention also provides an embodied intelligent agent, which may include: an embodied intelligent agent body and a processor. The processor executes a computer program to implement the embodied intelligent agent's fault-tolerant control method for warehouse handling. The embodied intelligent agent provided in this embodiment can comprehensively monitor and autonomously handle operational anomalies, thereby improving the reliability and continuity of operation of the embodied intelligent agent in complex dynamic environments.

[0112] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute an embodied intelligent agent storage and handling fault-tolerant control method. This method includes: real-time monitoring of the operating parameters of core operating components during the operation of the embodied intelligent agent, wherein the core operating components are those affecting the normal operation of the embodied intelligent agent; determining whether the core operating components are abnormal based on the operating parameters; determining the abnormality level of the embodied intelligent agent based on the abnormal core operating components when the core operating components are abnormal; determining an abnormality handling scheme matching the abnormality level of the embodied intelligent agent, and controlling the operation of the embodied intelligent agent based on the abnormality handling scheme.

[0113] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0114] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the embodied intelligent agent storage and handling fault-tolerant control method provided by the above methods. The method includes: real-time monitoring of the operating parameters of the core operating components of the embodied intelligent agent during operation, wherein the core operating components are components that affect the normal operation of the embodied intelligent agent; determining whether the core operating components are abnormal based on the operating parameters; determining the abnormality level of the embodied intelligent agent based on the abnormal core operating components when the core operating components are abnormal; determining an abnormality handling scheme matching the abnormality level according to the abnormality level of the embodied intelligent agent, and controlling the operation of the embodied intelligent agent based on the abnormality handling scheme.

[0115] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the embodied intelligent agent storage and handling fault-tolerant control method provided by the above methods. The method includes: real-time monitoring of the operating parameters of the core operating components of the embodied intelligent agent during operation, wherein the core operating components are components that affect the normal operation of the embodied intelligent agent; determining whether the core operating components are abnormal based on the operating parameters; determining the abnormality level of the embodied intelligent agent based on the abnormal core operating components when the core operating components are abnormal; determining an abnormality handling scheme matching the abnormality level according to the abnormality level of the embodied intelligent agent, and controlling the operation of the embodied intelligent agent based on the abnormality handling scheme.

[0116] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0117] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0118] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fault-tolerant control method for embodied intelligent agents in warehouse handling, characterized in that, The method includes: Real-time monitoring of the operating parameters of the core operating components during the operation of the embodied intelligent agent, wherein the core operating components are the components that affect the normal operation of the embodied intelligent agent; Based on the operating parameters, determine whether there is any abnormality in the core operating component; In the event of an anomaly in the core operating component, the anomaly level of the embodied intelligent agent is determined based on the anomaly in the core operating component. Based on the anomaly level of the embodied intelligent agent, an anomaly handling scheme matching the anomaly level is determined, and the operation of the embodied intelligent agent is controlled based on the anomaly handling scheme.

2. The embodied intelligent agent warehouse handling fault-tolerant control method according to claim 1, characterized in that, The step of determining whether the core operating component is abnormal based on the operating parameters includes: For each of the core operating components, determine the operating parameter thresholds of the core operating components under normal operating conditions; Based on the operating parameters and the operating parameter thresholds, it is determined whether the core operating component is abnormal. The operating parameter thresholds for the core operating component under normal operating conditions are determined using the following method: The system acquires multiple sets of historical operating parameters of the core operating components under normal operating conditions of the embodied intelligent agent, wherein the historical operating parameters are the historical operating parameters formed by the core operating components under a preset time period, and the preset time period is less than or equal to a time threshold. Based on multiple sets of historical operating parameters, the operating parameter thresholds of the core operating components under normal operating conditions are determined.

3. The embodied intelligent agent warehouse handling fault-tolerant control method according to claim 1, characterized in that, The determination of the anomaly level of the embodied intelligent agent based on the presence of abnormal core operating components includes: Determine the weight value of each of the core operating components, wherein the weight value is positively correlated with the degree of influence of the core operating component on the normal operation of the embodied intelligent agent; Assign an anomaly score to the core operating component that has an anomaly, and obtain the anomaly score of the core operating component that has an anomaly; Based on the anomaly scores of the core operating components that are abnormal, and the weight values ​​of the core operating components that are abnormal, the overall anomaly score of the embodied intelligent agent is determined. The abnormality level of the embodied agent is determined based on its comprehensive abnormality score.

4. The embodied intelligent agent warehouse handling fault-tolerant control method according to claim 3, characterized in that, The anomaly levels include mild anomaly, moderate anomaly, and severe anomaly; determining the anomaly level of the embodied agent based on its comprehensive anomaly score includes: If the overall anomaly score of the embodied agent is less than the first score threshold, the anomaly level of the embodied agent is determined to be the mild anomaly level. If the overall anomaly score of the embodied agent is greater than or equal to the first score threshold and less than the second score threshold, the anomaly level of the embodied agent is determined to be the moderate anomaly level. If the overall anomaly score of the embodied agent is greater than or equal to the second score threshold, the anomaly level of the embodied agent is determined to be the severe anomaly level.

5. The embodied intelligent agent warehouse handling fault-tolerant control method according to claim 4, characterized in that, The step of determining an anomaly handling scheme matching the anomaly level of the embodied intelligent agent includes: In response to the abnormality level of the embodied intelligent agent being the mild abnormality level, the abnormality handling scheme is determined to be a silent repair handling scheme, wherein the silent repair handling scheme is used to characterize the calibration of the core operating components with abnormality without interrupting the operation of the embodied intelligent agent; In response to the abnormality level of the embodied intelligent agent being the moderate abnormality level, the abnormality handling scheme is determined to be an active intervention handling scheme, wherein the active intervention handling scheme is used to characterize the calibration of the core operating components with abnormality when the operation of the embodied intelligent agent is interrupted; In response to the abnormality level of the embodied intelligent agent being the severe abnormality level, the abnormality handling scheme is determined to be a redundancy switching scheme, wherein the redundancy switching scheme is used to indicate that a backup embodied intelligent agent with the same operational capabilities as the embodied intelligent agent is invoked for operation.

6. The embodied intelligent agent warehouse handling fault-tolerant control method according to claim 5, characterized in that, The calibration of the abnormal core operating components without interrupting the operation of the embodied intelligent agent is achieved in the following manner: Without interrupting the operation of the embodied intelligent agent, the reference operating parameters of the core operating components that are malfunctioning are determined; Based on the reference operating parameters, the core operating components exhibiting abnormalities are calibrated, wherein the reference operating parameters are determined in the following manner: Obtain multiple sets of historical operating parameters of the core operating components of the embodied intelligent agent when performing the same task and operating normally, and determine the reference operating parameters based on the historical operating parameters, or Multiple sets of experimental operating parameters of the core operating components of the embodied intelligent agent are obtained when the agent performs the same task in the experimental environment and operates normally. Based on the experimental operating parameters, the reference operating parameters are determined.

7. The embodied intelligent agent warehouse handling fault-tolerant control method according to claim 5, characterized in that, The calibration of the abnormal core operating components when the operation of the embodied intelligent agent is interrupted includes: In the event that the operation of the embodied intelligent agent is interrupted, the core operating component is invoked to repeatedly execute the operating action within a preset number of times until the operating action meets the preset requirements, so as to calibrate the core operating component that has an abnormality.

8. The embodied intelligent agent warehouse handling fault-tolerant control method according to claim 7, characterized in that, The method further includes: If the operation results obtained after repeatedly executing the operation action by the core operating component within a preset number of times do not meet the preset requirements, a backup embodied intelligent agent with the same operational capabilities as the embodied intelligent agent is invoked to perform the operation.

9. A fault-tolerant control system for embodied intelligent agents in warehouse handling, characterized in that, The system includes: The monitoring module is used to monitor the operating parameters of the core operating components of the embodied intelligent agent in real time during operation, wherein the core operating components are the components that affect the normal operation of the embodied intelligent agent; The judgment module is used to determine whether there is any abnormality in the core operating component based on the operating parameters; The determination module is used to determine the anomaly level of the embodied intelligent agent based on the anomaly of the core operating component when the core operating component is abnormal. The processing module is used to determine an anomaly handling scheme that matches the anomaly level of the embodied intelligent agent, and control the operation of the embodied intelligent agent based on the anomaly handling scheme.

10. An embodied intelligent agent, characterized in that, The embodied intelligent agent includes: The embodied intelligent agent itself, and A processor that, when executing a computer program, implements the embodied intelligent agent warehouse handling fault-tolerant control method as described in any one of claims 1 to 8.

Images