Robot cluster cooperative control task allocation system based on internet of things

Abstract

The application discloses a robot cluster cooperative control task allocation system based on Internet of Things, and relates to the technical field of intelligent control of Internet of Things, and comprises an abnormality sensing unit, a traceability positioning unit, an identification quantization unit and a strategy matching unit, wherein: the abnormality sensing unit is used for acquiring local operation state parameters through a state sensor carried by a robot body, acquiring observation data of neighbor robots through a communication link between the robots, acquiring global environment reference data through an environment reference sensor fixedly arranged in a working environment, and constructing a multi-dimensional abnormal signal acquisition system covering a body layer, an interaction layer and an environment layer; the traceability positioning unit is used for extracting abnormal signals carried from the local operation state parameters, the neighbor robot observation data and the global environment reference data; the system calculation overhead and scheduling disturbance can be effectively reduced, and the scheduling efficiency, resource utilization rate and operation stability of the robot cluster in a dynamic disturbance environment can be significantly improved.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) intelligent control technology, specifically to an IoT-based robot swarm collaborative control task allocation system. Background Technology

[0002] In modern warehousing and logistics, intelligent manufacturing, and smart healthcare, robot swarm collaborative operation technology has been widely used. Multiple robots are interconnected through the Internet of Things (IoT) communication network to jointly complete complex tasks such as material handling, cargo sorting, and equipment inspection, effectively improving operational efficiency. However, during robot swarm collaborative operation, when a robot performing a task suddenly experiences a physical failure such as a burnt-out drive motor or stuck wheels, or encounters congestion due to obstacles blocking its path, the system often needs to promptly reassign the unfinished tasks of the faulty robot to other available robots to ensure the continuity of the overall operation.

[0003] In current robot swarm task allocation systems, when a robot malfunctions or encounters congestion, the system primarily relies on the robot's own heartbeat packets or fault codes to detect the anomaly. This lacks the ability to trace the source of the abnormal signal in time and space, making it impossible to distinguish between the actual fault and secondary abnormal signals caused by path obstruction. This can easily lead to the misjudgment of normal robots affected over a large area as the source of the fault, triggering a large-scale invalid task reassignment. Furthermore, because the existing system has a single source of anomaly detection, when a robot loses its heartbeat due to communication interference but its physical functions are normal, the system may misjudge it as a fault. When a robot's functions are only partially impaired (such as a robotic arm malfunction but normal movement), the system cannot identify its remaining operational capabilities and can only shut it down entirely, resulting in wasted resources and hindering its usability.

[0004] Therefore, in view of this, the present invention proposes a robot swarm collaborative control task allocation system based on the Internet of Things to make up for and improve the shortcomings of the existing technology. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides an Internet of Things-based robot swarm collaborative control task allocation system, thereby resolving the corresponding technical issues raised in the background section.

[0006] To achieve the above objectives, the technical solution adopted by this invention is: a robot swarm collaborative control task allocation system based on the Internet of Things, comprising an anomaly perception unit, a source tracing and localization unit, an identification and quantification unit, and a strategy matching unit, wherein: The anomaly perception unit is used to acquire the robot's own operating status parameters through the status sensors mounted on the robot body, acquire the observation data of neighboring robots through the communication link between robots, and acquire global environmental reference data through environmental reference sensors fixedly deployed in the working environment, thus constructing a multi-dimensional anomaly signal acquisition system covering the body layer, interaction layer and environment layer. The source tracing and localization unit is used to extract the abnormal signals carried from the local robot's operating status parameters, neighbor robot observation data, and global environment reference data. Using the preset time window length and spatial radius length as clustering parameters, it performs spatiotemporal clustering on the extracted abnormal signals to divide them into abnormal signal clusters. It extracts the spatial coordinates corresponding to the abnormal signal with the earliest trigger timestamp in each abnormal signal cluster, determines the spatial coordinates as the source coordinates of the abnormal signal cluster, and determines the source type of the abnormal signal cluster based on the local robot's operating status parameters corresponding to the source coordinates. The identification and quantization unit is used to classify the fault into one of four preset levels when it is determined to be a primary fault of the robot body, based on the subsystem state feedback data of the robot: complete failure level, loss of motion function level, degradation of operation function level, and communication and sensing isolation level. It obtains the support of three sources for the divided fault level from the local operating status parameters, neighbor robot observation data, and global environment reference data, counts the number of sources that support the fault level determination, and determines the confidence score of the determination result based on the number of sources. The strategy matching unit is used to match the corresponding handling action from the preset strategy library according to the source type, fault level and confidence score, and then issue it for execution.

[0007] As a preferred approach, the specific process for constructing a multi-dimensional anomaly signal acquisition system covering the ontology layer, interaction layer, and environment layer is as follows: The robot body is equipped with a drive motor current sensor, inertial measurement unit, battery voltage monitoring module and robotic arm joint angle feedback module to collect real-time operating status parameters. These parameters include drive motor current value, body attitude angle, battery voltage value and robotic arm joint angle deviation value, which constitute the body layer data. The robot receives observation data broadcast by neighboring robots through a communication link between robots. The observation data includes the robot's position coordinates, movement speed, and distance between the robot and obstacles in front of it as observed by neighboring robots. At the same time, the robot obtains physical detection data of the environment in front of it through its onboard lidar, forming the interaction layer data. The global spatial coordinates of each robot are obtained by UWB positioning base stations that are fixedly deployed in the working environment, and the real-time pose image data of each robot is obtained by monitoring cameras that are fixedly deployed in the working environment, thus forming environmental layer data.

[0008] As a preferred approach, the specific process of performing spatiotemporal clustering on the extracted anomalous signals to divide them into anomalous signal clusters is as follows: The abnormal signals carried by the machine's operating status parameters, neighbor robot observation data and global environment reference data are extracted. Each abnormal signal carries the source robot identifier, abnormality type code, trigger timestamp and trigger spatial coordinates. Using the preset time window length and spatial radius length as clustering parameters, density-based spatiotemporal clustering is performed on all extracted abnormal signals, and abnormal signals whose trigger timestamps fall within the same time window and whose trigger spatial coordinates fall within the same spatial radius are classified into the same abnormal signal cluster. Traverse all abnormal signal clusters, extract the spatial coordinates corresponding to the abnormal signal with the earliest trigger timestamp in each abnormal signal cluster, and determine the spatial coordinates as the source coordinates of the abnormal signal cluster.

[0009] As a preferred method, the specific process for determining the source type of the abnormal signal cluster based on the robot's on-premises operating status parameters at the source coordinates is as follows: Obtain the robot's local operating status parameters corresponding to the origin coordinates, and determine whether there is an abnormal jump in the drive motor current value in the local operating status parameters. The abnormal jump refers to the drive motor current value returning to zero or the drive motor current value continuously exceeding the rated current value. When there is an abnormal jump in the current value of the robot's drive motor at the origin coordinate, the source type of the abnormal signal cluster is determined to be a primary fault of the robot body. When there is no abnormal jump in the current value of the robot's drive motor at the source coordinate, and the abnormal type code of the abnormal signal in the abnormal signal cluster is forward ranging abnormal, the source type of the abnormal signal cluster is determined to be path congestion. When the abnormal signal cluster contains only abnormal signals from the same robot, and there is no abnormal jump in the current value of the robot's drive motor, the source type of the abnormal signal cluster is determined to be a sensor false alarm or a momentary interruption of the communication link.

[0010] As a preferred option, the specific process for classifying faults into one of the four preset levels—complete failure level, loss of motion function level, operational function degradation level, and communication / sensing isolation level—is as follows: Obtain the subsystem status feedback data of the robot, which includes the drive motor current feedback value, the robotic arm joint angle deviation feedback value, and the online status value of the communication module; When the drive motor current feedback value is zero and the online status value of the communication module is offline, the fault is classified as a complete failure layer. When the current feedback value of the drive motor is continuously greater than the rated current value and the online status value of the communication module is online, the fault is classified as the loss of motion function. When the drive motor current feedback value is less than or equal to the rated current value, or the robotic arm joint angle deviation feedback value is continuously greater than the preset deviation threshold, the fault is classified as the operational function degradation layer. When the drive motor current feedback value is less than or equal to the rated current value, the robotic arm joint angle deviation feedback value is less than or equal to the preset deviation threshold, the online status value of the communication module is offline, or the data output of the on-board sensor is abnormal, the fault is classified as the communication and sensing isolation layer.

[0011] As a preferred approach, the specific process of determining the confidence score of the judgment result based on the number of sources supporting the fault level determination among the three sources is as follows: Obtain the support status of the machine's operating status parameters for the identified fault levels: When the drive motor current feedback value or the robotic arm joint angle deviation feedback value in the machine's operating status parameters matches the fault level determination conclusion, confirm that the body layer supports the determination. Obtain the support of neighbor robot observation data for the identified fault levels: when the position coordinates and movement speed of the robot observed by the neighbor robot match the robot's movement state, confirm that the interaction layer supports this determination; Obtain the support of global environment reference data for the identified fault levels: When the global spatial coordinates of the robot obtained by the UWB positioning base station or the real-time pose image data of the robot obtained by the monitoring camera matches the robot's motion state, it is confirmed that the environment layer supports the determination. The number of sources in the ontology layer, interaction layer, and environment layer that support the determination of the fault level is counted. When all three sources support the determination of the fault level, the output confidence score is the first threshold; when two sources support the determination of the fault level, the output confidence score is the second threshold; when one source supports the determination of the fault level, the output confidence score is the third threshold. Wherein, the first threshold is greater than the second threshold, and the second threshold is greater than the third threshold.

[0012] As a preferred option, the specific process of matching the corresponding handling action from a preset strategy library based on the source type, fault level, and confidence score is as follows: Obtain the source type, fault level, and confidence score, and compare them based on a preset matching threshold as follows: When the source type is a primary fault of the robot body, the fault level is a complete failure level, and the confidence score is greater than or equal to the first matching threshold, all unfinished tasks of the robot are released to the global task pool, triggering global reallocation. When the source type is a primary fault of the robot body, the fault level is the loss of motion function, and the confidence score is greater than or equal to the second matching threshold, the tasks in the robot that depend on motion function are transferred to other robots, and the tasks in the robot that do not depend on motion function are suspended. When the source type is a primary fault of the robot body, the fault level is the operational function degradation layer, and the confidence score is greater than or equal to the third matching threshold, the robot is marked as functionally limited, and subsequent task types that match the remaining functions are only assigned to the robot. When the source type is a primary fault of the robot body, the fault level is the communication and sensing isolation layer, and the confidence score is greater than or equal to the fourth matching threshold, a reset command is issued to the robot without triggering task transfer. When the source type is path congestion, a local path replanning instruction is issued to the affected robots without changing the task assignment of any robot. When the source type is a sensor false alarm or a momentary interruption of the communication link, a calibration command or reset command is issued to the robot, and the corresponding sensor data of the robot is temporarily downgraded without triggering a task transfer.

[0013] Compared with existing technologies, the beneficial effects of this invention are as follows: By acquiring local operating status parameters, observation data from neighboring robots, and global environmental reference data, a multi-dimensional abnormal signal acquisition system covering the body layer, interaction layer, and environment layer is constructed. Abnormal signals are extracted and spatiotemporally clustered to divide abnormal signal clusters, determine source coordinates, and identify the source type. When a primary fault of the robot body is identified, the fault level is classified based on the robot's subsystem state feedback data. The support for the classified fault level is obtained from the local operating status parameters, neighboring robot observation data, and global environmental reference data, and the support from these three sources is statistically analyzed to determine the fault level. The system identifies the number of sources and determines the confidence score of the judgment result based on the number of sources. According to the source type, fault level, and confidence score, it matches the corresponding handling action from the preset strategy library and issues it for execution. It can distinguish between primary faults and secondary impact signals to avoid invalid reallocation, identify the degradation state of some robot functions and degrade them to improve resource utilization, identify sensor false alarms through multi-source cross-verification to avoid single-point anomalies causing system malfunctions, and adopt graded handling strategies according to different anomaly types and confidence levels. This effectively reduces system computational overhead and scheduling disturbances, and significantly improves the scheduling efficiency, resource utilization, and operational stability of robot clusters in dynamic disturbance environments. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall structure of a preferred embodiment of the present invention; Figure 2This is a system flowchart of the present invention; Figure 3 This is a flowchart illustrating the process of dividing abnormal signal clusters according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Embodiments of the present invention: Please refer to Figures 1 to 3 As shown, the IoT-based robot swarm collaborative control task allocation system includes an anomaly perception unit, a source tracing and localization unit, an identification and quantification unit, and a strategy matching unit, wherein: The anomaly perception unit is used to acquire the robot's own operating status parameters through the status sensors mounted on the robot body, acquire the observation data of neighboring robots through the communication link between robots, and acquire global environmental reference data through environmental reference sensors fixedly deployed in the working environment, thus constructing a multi-dimensional anomaly signal acquisition system covering the body layer, interaction layer and environment layer. The source tracing and localization unit is used to extract the abnormal signals carried from the local robot's operating status parameters, neighbor robot observation data, and global environment reference data. Using the preset time window length and spatial radius length as clustering parameters, it performs spatiotemporal clustering on the extracted abnormal signals to divide them into abnormal signal clusters. It extracts the spatial coordinates corresponding to the abnormal signal with the earliest trigger timestamp in each abnormal signal cluster, determines the spatial coordinates as the source coordinates of the abnormal signal cluster, and determines the source type of the abnormal signal cluster based on the local robot's operating status parameters corresponding to the source coordinates. The identification and quantization unit is used to classify the fault into one of four preset levels when it is determined to be a primary fault of the robot body, based on the subsystem state feedback data of the robot: complete failure level, loss of motion function level, degradation of operation function level, and communication and sensing isolation level. It obtains the support of three sources for the divided fault level from the local operating status parameters, neighbor robot observation data, and global environment reference data, counts the number of sources that support the fault level determination, and determines the confidence score of the determination result based on the number of sources. The strategy matching unit is used to match the corresponding handling action from the preset strategy library according to the source type, fault level and confidence score, and then issue it for execution.

[0017] The specific process of constructing a multi-dimensional anomaly signal acquisition system covering the ontology layer, interaction layer, and environment layer is as follows: The robot body is equipped with a drive motor current sensor, inertial measurement unit, battery voltage monitoring module and robotic arm joint angle feedback module to collect real-time operating status parameters. These parameters include drive motor current value, body attitude angle, battery voltage value and robotic arm joint angle deviation value, which constitute the body layer data. The robot receives observation data broadcast by neighboring robots through a communication link between robots. The observation data includes the robot's position coordinates, movement speed, and distance between the robot and obstacles in front of it as observed by neighboring robots. At the same time, the robot obtains physical detection data of the environment in front of it through its onboard lidar, forming the interaction layer data. The global spatial coordinates of each robot are obtained by UWB positioning base stations that are fixedly deployed in the working environment, and the real-time pose image data of each robot is obtained by monitoring cameras that are fixedly deployed in the working environment, thus forming environmental layer data.

[0018] The specific process of performing spatiotemporal clustering on the extracted anomalous signals to divide them into anomalous signal clusters is as follows: The abnormal signals carried by the machine's operating status parameters, neighbor robot observation data and global environment reference data are extracted. Each abnormal signal carries the source robot identifier, abnormality type code, trigger timestamp and trigger spatial coordinates. Using the preset time window length and spatial radius length as clustering parameters, density-based spatiotemporal clustering is performed on all extracted abnormal signals, and abnormal signals whose trigger timestamps fall within the same time window and whose trigger spatial coordinates fall within the same spatial radius are classified into the same abnormal signal cluster. Traverse all abnormal signal clusters, extract the spatial coordinates corresponding to the abnormal signal with the earliest trigger timestamp in each abnormal signal cluster, and determine the spatial coordinates as the source coordinates of the abnormal signal cluster.

[0019] The specific process for determining the source type of this abnormal signal cluster based on the robot's on-premises operating status parameters at the origin coordinates is as follows: Obtain the robot's local operating status parameters corresponding to the origin coordinates, and determine whether there is an abnormal jump in the drive motor current value in the local operating status parameters. The abnormal jump refers to the drive motor current value returning to zero or the drive motor current value continuously exceeding the rated current value. When there is an abnormal jump in the current value of the robot's drive motor at the origin coordinate, the source type of the abnormal signal cluster is determined to be a primary fault of the robot body. When there is no abnormal jump in the current value of the robot's drive motor at the source coordinate, and the abnormal type code of the abnormal signal in the abnormal signal cluster is forward ranging abnormal, the source type of the abnormal signal cluster is determined to be path congestion. When the abnormal signal cluster contains only abnormal signals from the same robot, and there is no abnormal jump in the current value of the robot's drive motor, the source type of the abnormal signal cluster is determined to be a sensor false alarm or a momentary interruption of the communication link.

[0020] The specific process of classifying faults into one of the four preset levels—complete failure level, loss of motion function level, operational function degradation level, and communication / sensor isolation level—is as follows: Obtain the subsystem status feedback data of the robot, which includes the drive motor current feedback value, the robotic arm joint angle deviation feedback value, and the online status value of the communication module; When the drive motor current feedback value is zero and the online status value of the communication module is offline, the fault is classified as a complete failure layer. When the current feedback value of the drive motor is continuously greater than the rated current value and the online status value of the communication module is online, the fault is classified as the loss of motion function. When the drive motor current feedback value is less than or equal to the rated current value, or the robotic arm joint angle deviation feedback value is continuously greater than the preset deviation threshold, the fault is classified as the operational function degradation layer. When the drive motor current feedback value is less than or equal to the rated current value, the robotic arm joint angle deviation feedback value is less than or equal to the preset deviation threshold, the online status value of the communication module is offline, or the data output of the on-board sensor is abnormal, the fault is classified as the communication and sensing isolation layer.

[0021] The specific process for determining the confidence score of the judgment result based on the number of sources supporting the fault level determination from the three sources is as follows: Obtain the support status of the machine's operating status parameters for the identified fault levels: When the drive motor current feedback value or the robotic arm joint angle deviation feedback value in the machine's operating status parameters matches the fault level determination conclusion, confirm that the body layer supports the determination. Obtain the support of neighbor robot observation data for the identified fault levels: when the position coordinates and movement speed of the robot observed by the neighbor robot match the robot's movement state, confirm that the interaction layer supports this determination; Obtain the support of global environment reference data for the identified fault levels: When the global spatial coordinates of the robot obtained by the UWB positioning base station or the real-time pose image data of the robot obtained by the monitoring camera matches the robot's motion state, it is confirmed that the environment layer supports the determination. The number of sources in the ontology layer, interaction layer, and environment layer that support the determination of the fault level is counted. When all three sources support the determination of the fault level, the output confidence score is the first threshold; when two sources support the determination of the fault level, the output confidence score is the second threshold; when one source supports the determination of the fault level, the output confidence score is the third threshold. Wherein, the first threshold is greater than the second threshold, and the second threshold is greater than the third threshold.

[0022] The specific process of matching the corresponding handling action from the preset strategy library based on the source type, fault level, and confidence score is as follows: Obtain the source type, fault level, and confidence score, and compare them based on a preset matching threshold as follows: When the source type is a primary fault of the robot body, the fault level is a complete failure level, and the confidence score is greater than or equal to the first matching threshold, all unfinished tasks of the robot are released to the global task pool, triggering global reallocation. When the source type is a primary fault of the robot body, the fault level is the loss of motion function, and the confidence score is greater than or equal to the second matching threshold, the tasks in the robot that depend on motion function are transferred to other robots, and the tasks in the robot that do not depend on motion function are suspended. When the source type is a primary fault of the robot body, the fault level is the operational function degradation layer, and the confidence score is greater than or equal to the third matching threshold, the robot is marked as functionally limited, and subsequent task types that match the remaining functions are only assigned to the robot. When the source type is a primary fault of the robot body, the fault level is the communication and sensing isolation layer, and the confidence score is greater than or equal to the fourth matching threshold, a reset command is issued to the robot without triggering task transfer. When the source type is path congestion, a local path replanning instruction is issued to the affected robots without changing the task assignment of any robot. When the source type is a sensor false alarm or a momentary interruption of the communication link, a calibration command or reset command is issued to the robot, and the corresponding sensor data of the robot is temporarily downgraded without triggering a task transfer.

[0023] By acquiring local operating status parameters, observation data from neighboring robots, and global environmental reference data, a multi-dimensional anomaly signal acquisition system covering the body layer, interaction layer, and environment layer is constructed. Anomaly signals are extracted and spatiotemporally clustered to classify anomaly signal clusters, determine source coordinates, and identify source types. When a primary fault of the robot body is identified, the fault level is classified based on the robot's subsystem state feedback data. The support of local operating status parameters, neighboring robot observation data, and global environmental reference data for the classified fault level is assessed, and the number of sources supporting the fault level determination is counted. The system determines the confidence score of the judgment result, matches the corresponding handling action from the preset strategy library according to the source type, fault level and confidence score, and issues the action for execution. It can distinguish between primary faults and secondary impact signals to avoid invalid reallocation, identify the degradation state of some robot functions and degrade them to improve resource utilization, identify sensor false alarms through multi-source cross-validation to avoid single-point anomalies causing system malfunctions, and adopt graded handling strategies according to different anomaly types and confidence levels. This effectively reduces system computational overhead and scheduling disturbances, and significantly improves the scheduling efficiency, resource utilization and operational stability of robot clusters in dynamic disturbance environments.

[0024] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0025] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation. In the two embodiments provided in this application, it should be understood that the disclosed apparatus and system can be implemented in other ways; for example, the apparatus embodiments described above are merely illustrative, and the division of modules is merely a logical functional division. In actual implementation, there may be other division methods, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed; furthermore, the coupling or direct coupling or communication connection between the shown or discussed mutuals can be through some interfaces, and the indirect coupling or communication connection between the apparatus or modules can be electrical, mechanical or other forms. The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A robot swarm collaborative control task allocation system based on the Internet of Things, characterized in that, It includes an anomaly detection unit, a source tracing and localization unit, an identification and quantification unit, and a policy matching unit, wherein: The anomaly perception unit is used to acquire the robot's own operating status parameters through the status sensors mounted on the robot body, acquire the observation data of neighboring robots through the communication link between robots, and acquire global environmental reference data through environmental reference sensors fixedly deployed in the working environment, thus constructing a multi-dimensional anomaly signal acquisition system covering the body layer, interaction layer and environment layer. The source tracing and localization unit is used to extract the abnormal signals carried from the local robot's operating status parameters, neighbor robot observation data, and global environment reference data. Using the preset time window length and spatial radius length as clustering parameters, it performs spatiotemporal clustering on the extracted abnormal signals to divide them into abnormal signal clusters. It extracts the spatial coordinates corresponding to the abnormal signal with the earliest trigger timestamp in each abnormal signal cluster, determines the spatial coordinates as the source coordinates of the abnormal signal cluster, and determines the source type of the abnormal signal cluster based on the local robot's operating status parameters corresponding to the source coordinates. The identification and quantization unit is used to classify the fault into one of four preset levels when it is determined to be a primary fault of the robot body, based on the subsystem state feedback data of the robot: complete failure level, loss of motion function level, degradation of operation function level, and communication and sensing isolation level. It obtains the support of three sources for the divided fault level from the local operating status parameters, neighbor robot observation data, and global environment reference data, counts the number of sources that support the fault level determination, and determines the confidence score of the determination result based on the number of sources. The strategy matching unit is used to match the corresponding handling action from the preset strategy library according to the source type, fault level and confidence score, and then issue it for execution.

2. The IoT-based robot swarm collaborative control task allocation system according to claim 1, characterized in that, The specific process of constructing a multi-dimensional anomaly signal acquisition system covering the ontology layer, interaction layer, and environment layer is as follows: The robot body is equipped with a drive motor current sensor, inertial measurement unit, battery voltage monitoring module and robotic arm joint angle feedback module to collect real-time operating status parameters. These parameters include drive motor current value, body attitude angle, battery voltage value and robotic arm joint angle deviation value, which constitute the body layer data. The robot receives observation data broadcast by neighboring robots through a communication link between robots. The observation data includes the robot's position coordinates, movement speed, and distance between the robot and obstacles in front of it as observed by neighboring robots. At the same time, the robot obtains physical detection data of the environment in front of it through its onboard lidar, forming the interaction layer data. The global spatial coordinates of each robot are obtained by UWB positioning base stations that are fixedly deployed in the working environment, and the real-time pose image data of each robot is obtained by monitoring cameras that are fixedly deployed in the working environment, thus forming environmental layer data.

3. The IoT-based robot swarm collaborative control task allocation system according to claim 1, characterized in that, The specific process of performing spatiotemporal clustering on the extracted anomalous signals to divide them into anomalous signal clusters is as follows: The abnormal signals carried by the machine's operating status parameters, neighbor robot observation data and global environment reference data are extracted. Each abnormal signal carries the source robot identifier, abnormality type code, trigger timestamp and trigger spatial coordinates. Using the preset time window length and spatial radius length as clustering parameters, density-based spatiotemporal clustering is performed on all extracted abnormal signals, and abnormal signals whose trigger timestamps fall within the same time window and whose trigger spatial coordinates fall within the same spatial radius are classified into the same abnormal signal cluster. Traverse all abnormal signal clusters, extract the spatial coordinates corresponding to the abnormal signal with the earliest trigger timestamp in each abnormal signal cluster, and determine the spatial coordinates as the source coordinates of the abnormal signal cluster.

4. The IoT-based robot swarm collaborative control task allocation system according to claim 1, characterized in that, The specific process for determining the source type of this abnormal signal cluster based on the robot's on-premises operating status parameters at the origin coordinates is as follows: Obtain the robot's local operating status parameters corresponding to the origin coordinates, and determine whether there is an abnormal jump in the drive motor current value in the local operating status parameters. The abnormal jump refers to the drive motor current value returning to zero or the drive motor current value continuously exceeding the rated current value. When there is an abnormal jump in the current value of the robot's drive motor at the origin coordinate, the source type of the abnormal signal cluster is determined to be a primary fault of the robot body. When there is no abnormal jump in the current value of the robot's drive motor at the source coordinate, and the abnormal type code of the abnormal signal in the abnormal signal cluster is forward ranging abnormal, the source type of the abnormal signal cluster is determined to be path congestion. When the abnormal signal cluster contains only abnormal signals from the same robot, and there is no abnormal jump in the current value of the robot's drive motor, the source type of the abnormal signal cluster is determined to be a sensor false alarm or a momentary interruption of the communication link.

5. The IoT-based robot swarm collaborative control task allocation system according to claim 1, characterized in that, The specific process of classifying faults into one of the four preset levels—complete failure level, loss of motion function level, operational function degradation level, and communication / sensor isolation level—is as follows: Obtain the subsystem status feedback data of the robot, which includes the drive motor current feedback value, the robotic arm joint angle deviation feedback value, and the online status value of the communication module; When the drive motor current feedback value is zero and the online status value of the communication module is offline, the fault is classified as a complete failure layer. When the current feedback value of the drive motor is continuously greater than the rated current value and the online status value of the communication module is online, the fault is classified as the loss of motion function. When the drive motor current feedback value is less than or equal to the rated current value, or the robotic arm joint angle deviation feedback value is continuously greater than the preset deviation threshold, the fault is classified as the operational function degradation layer. When the drive motor current feedback value is less than or equal to the rated current value, the robotic arm joint angle deviation feedback value is less than or equal to the preset deviation threshold, the online status value of the communication module is offline, or the data output of the on-board sensor is abnormal, the fault is classified as the communication and sensing isolation layer.

6. The IoT-based robot swarm collaborative control task allocation system according to claim 1, characterized in that, The specific process for determining the confidence score of the judgment result based on the number of sources supporting the fault level determination from the three sources is as follows: Obtain the support status of the machine's operating status parameters for the identified fault levels: When the drive motor current feedback value or the robotic arm joint angle deviation feedback value in the machine's operating status parameters matches the fault level determination conclusion, confirm that the body layer supports the determination. Obtain the support of neighbor robot observation data for the identified fault levels: when the position coordinates and movement speed of the robot observed by the neighbor robot match the robot's movement state, confirm that the interaction layer supports this determination; Obtain the support of global environment reference data for the identified fault levels: When the global spatial coordinates of the robot obtained by the UWB positioning base station or the real-time pose image data of the robot obtained by the monitoring camera matches the robot's motion state, it is confirmed that the environment layer supports the determination. The number of sources in the ontology layer, interaction layer, and environment layer that support the determination of the fault level is counted. When all three sources support the determination of the fault level, the output confidence score is the first threshold. When two sources support the determination of the fault level, the output confidence score is the second threshold; when one source supports the determination of the fault level, the output confidence score is the third threshold. Wherein, the first threshold is greater than the second threshold, and the second threshold is greater than the third threshold.

7. The IoT-based robot swarm collaborative control task allocation system according to claim 1, characterized in that, The specific process of matching the corresponding handling action from the preset strategy library based on the source type, fault level, and confidence score is as follows: Obtain the source type, fault level, and confidence score, and compare them based on a preset matching threshold as follows: When the source type is a primary fault of the robot body, the fault level is a complete failure level, and the confidence score is greater than or equal to the first matching threshold, all unfinished tasks of the robot are released to the global task pool, triggering global reallocation. When the source type is a primary fault of the robot body, the fault level is the loss of motion function, and the confidence score is greater than or equal to the second matching threshold, the tasks in the robot that depend on motion function are transferred to other robots, and the tasks in the robot that do not depend on motion function are suspended. When the source type is a primary fault of the robot body, the fault level is the operational function degradation layer, and the confidence score is greater than or equal to the third matching threshold, the robot is marked as functionally limited, and subsequent task types that match the remaining functions are only assigned to the robot. When the source type is a primary fault of the robot body, the fault level is the communication and sensing isolation layer, and the confidence score is greater than or equal to the fourth matching threshold, a reset command is issued to the robot without triggering task transfer. When the source type is path congestion, a local path replanning instruction is issued to the affected robots without changing the task assignment of any robot. When the source type is a sensor false alarm or a momentary interruption of the communication link, a calibration command or reset command is issued to the robot, and the corresponding sensor data of the robot is temporarily downgraded without triggering a task transfer.

Images