Robot node allocation method and system

Abstract

A robot node scheduling method and system are disclosed. The robot node scheduling method includes: establishing a network structure containing multiple connectors; when multiple telemetry data from the multiple connectors meet preset conditions, obtaining the position information of a mobile vehicle from one of the connectors, using the one connector as a starting point, determining another connector as a first target point, obtaining the shortest path from the starting point to the first target point as a first forwarding path, transmitting data from the mobile vehicle to the first target point according to the first forwarding path, determining a second forwarding path based at least on the multiple telemetry data, position information, speed information of the mobile vehicle, and scheduling decision parameters, and transmitting data from the first target point to the target point of the second forwarding path according to the second forwarding path.

Description

Technical Field

[0001] This invention relates to a robot node scheduling method and system. Prior Technology

[0002] The application of the new generation of Robot Operating System 2 (ROS2) is growing rapidly. The Internet of Robots requires network capacity and quality of service (QoS) to meet the needs of users in the context of a data distribution service (DDS) global network operation environment.

[0003] Because distributed data service architectures use broadcasting to transmit messages, resulting in large packet volumes, distributed data services cannot operate smoothly over the Internet. Second-generation robot operation systems can deploy nodes to edge computing devices to meet the high computational demands, latency sensitivity, and stringent quality-of-service requirements of robotic applications. However, current network technology limits the flexibility of dynamically adjusting second-generation robot operation system nodes, thus failing to effectively reduce computational latency. Summary of the Invention

[0004] In view of the above, the present invention provides a robot node scheduling method and system for solving the above problems.

[0005] A robot node scheduling method according to an embodiment of the present invention includes the following steps, performed by at least one processing device: establishing a network structure, wherein the network structure includes multiple connectors corresponding to multiple work areas; acquiring multiple telemetry data from the multiple connectors and determining whether the multiple telemetry data meet preset conditions; generating a scheduling trigger command when the multiple telemetry data meet the preset conditions or when a push message is generated according to a user instruction; responding to the scheduling trigger command, acquiring scheduling decision parameters from an operation control unit and performing node mapping, wherein the scheduling decision parameters indicate a first target edge computing device, and the node mapping includes: acquiring the position information of a mobile vehicle from one of the multiple connectors and using the one of the multiple connectors as a starting point; determining another of the multiple connectors as a first target point according to the network structure and the scheduling decision parameters, wherein the first target point corresponds to the first target edge computing device; acquiring the shortest path from the starting point to the first target point according to the network structure, as a first forwarding path; and deploying a second-generation Robot Operating System. 2. ROS2) node to the first target edge computing device; transmit data from the mobile vehicle to the first target point according to the first forwarding path; execute the scheduling reverse forwarding algorithm to determine the second forwarding path based at least according to the multiple telemetry data, location information, speed information of the mobile vehicle and scheduling decision parameters, wherein the scheduling reverse forwarding algorithm is based on Markov decision process; and transmit data from the first target point to the target point indicated by the second forwarding path according to the second forwarding path.

[0006] A robot node scheduling system according to an embodiment of the present invention includes: at least one memory and at least one processing device. The at least one memory stores a plurality of instructions. The at least one processing device is connected to the at least one memory and is used to execute the plurality of instructions to perform the robot node scheduling method as described above.

[0007] In summary, the robot node scheduling method and system according to one or more embodiments of the present invention provide scheduling for second-generation nodes of robot operating systems in a robot internet under the background of mobile edge computing. Based on network topology and the location information of mobile vehicles, it performs data transmission routing planning, not only planning the scheduling from mobile vehicles to edge computing devices, but also considering the movement status of mobile vehicles during the data transmission process from edge computing devices to mobile vehicles. Therefore, the robot node scheduling method and system according to one or more embodiments of the present invention can manage each connector in the network structure in a flexible and simplified manner, thereby reducing the latency of computing tasks in the robot internet and improving the overall network performance.

[0008] The foregoing description of the disclosure and the following description of the embodiments are intended to demonstrate and explain the spirit and principles of the present invention, and to provide a further explanation of the scope of the patent application of the present invention. Simple Explanation of the Diagram

[0009] Figure 1 is a block diagram of a robot node scheduling system according to an embodiment of the present invention. Figure 2 is an architecture diagram of the deployment architecture of a robot node scheduling system according to an embodiment of the present invention. Figure 3 is a flowchart illustrating a robot node scheduling method according to an embodiment of the present invention. Figure 4 is a schematic diagram of a network structure according to an embodiment of the present invention. Figure 5 is a flowchart illustrating the establishment of a network structure in a robot node scheduling method according to an embodiment of the present invention. Figure 6 is a flowchart illustrating the robot node scheduling method for obtaining the first forwarding path according to an embodiment of the present invention. Implementation

[0010] The following detailed description of the features and advantages of the present invention is sufficient to enable anyone skilled in the art to understand the technical content of the present invention and implement it accordingly. Based on the disclosure, patent claims, and drawings in this specification, anyone skilled in the art can easily understand the relevant objectives and advantages of the present invention. The following embodiments further illustrate the points of the present invention, but are not intended to limit the scope of the present invention in any way.

[0011] The robot node scheduling system and method described below, representing one or more embodiments, are applicable to the working environment of the Internet of Robots. For example, when the working environment is a factory, a work area may include at least one mobile vehicle (such as a robot).

[0012] Please refer to Figure 1, which is a block diagram of a robot node scheduling system according to an embodiment of the present invention. As shown in Figure 1, the robot node scheduling system 1 includes at least one memory 11 and at least one processing device 12. The processing device 12 is electrically or communicatively connected to the memory 11.

[0013] Memory 11 is used to store multiple instructions, which may correspond to robot node scheduling methods. Memory 11 may be non-volatile memory (NVM), such as flash memory, non-volatile random access memory (NVRAM), etc. Memory 11 can store network structure model data, especially global network structure model data.

[0014] The processing device 12 is used to execute the plurality of instructions stored in the memory 11 to perform one or more embodiments of the robot node scheduling method described below. Further, in embodiments where the number of processing devices 12 is one, the processing device 12 can be used to execute all steps of the robot node scheduling method described below; in embodiments where the number of processing devices 12 is multiple, each of the multiple processing devices 12 can be used to execute one or more steps of the robot node scheduling method described below. The processing device 12 may include one or more processors, such as a central processing unit, graphics processor, microcontroller, programmable logic controller, or other processor with signal processing capabilities.

[0015] Please refer to Figure 2, which is an architecture diagram of the deployment architecture of a robot node scheduling system according to an embodiment of the present invention. As shown in Figure 2, the deployment architecture 2 includes a connector 20, a network modeler 21, a global network model storage area 22, a quality of service (QoS) sensor 23, a mapper 24, a network conditioner 25, and a Robot Operating System 2 (ROS2) launcher 26. The deployment architecture 2 can interact with the database A1 and the operation control unit A2. The database A1 can be used to store the working environment map and network structure, and can also store preset conditions related to QoS. The operation control unit A2 can be implemented using a Kubernetes (K8) system and includes a scheduling strategy and deployment functions, wherein the scheduling strategy is used to obtain the optimal scheduling decision parameters. The scheduling decision parameters indicate an edge computing device in the network structure, especially the edge computing device most suitable for processing the current task.

[0016] The second-generation connector 20 of the robot operating system is connected to the network modeler 21, the service quality sensor 23, and the network conditioner 25; the network modeler 21 is further connected to the global network model storage area 22 and the database A1; the global network model storage area 22 is further connected to the service quality sensor 23, the network conditioner 25, the second-generation robot operating system launcher 26, and the database A1; the service quality sensor 23 is further connected to the network conditioner 25 and the second-generation robot operating system launcher 26; the mapper 24 is connected to the network conditioner 25, the second-generation robot operating system launcher 26, and the operation control unit A2; the second-generation robot operating system launcher 26 is further connected to the database A1. The second-generation robot operating system launcher 26 can be further connected to the user interface. It should be noted that, as illustrated in Figure 2, communication between functional blocks is bidirectional. However, in some embodiments, the communication path between some functional blocks can be unidirectional and is not limited to bidirectional.

[0017] The network modeler 21 can obtain a working environment map from the database A1, and based on this, establish a network structure composed of multiple connectors 20, which is then mapped to the global network model storage area 22. The network modeler 21 obtains the network structure's configuration data based on telemetry data obtained from the connectors 20, and merges this configuration data into the global network model storage area 22. This network structure is specifically a global network structure. Furthermore, the network structure stored in the global network model storage area 22 can also be recorded in the database A1. Thus, when the system restarts, the second-generation robot operating system ejector 26 can import the network structure recorded in the database A1 into the global network model storage area 22. Figure 2 illustrates an example of a connector 20, but the deployment architecture 2 may include multiple connectors 20, each corresponding to a different work area, each connected to the network modeler 21, the quality of service sensor 23, and the network conditioner 25, and one of the multiple connectors 20 may be connected to one or more of the remaining connectors 20.

[0018] Connector 20 can obtain data from mobile vehicles and / or computing devices through its network domain and can forward the data to another connector 20. Connector 20 can rely on Data Distributed Service Discovery (DDS Discovery) technology to execute various connections conforming to the Data Distributed Service communication protocol. Connector 20 can control path switching according to the flow rules planned by network conditioner 25. Connector 20 can monitor to collect telemetry data and provide it to network modeler 21 and / or quality of service (QoS) sensor 23. QoS sensor 23 can use the telemetry data from connector 20 to determine whether to trigger node scheduling of Robot Operating System 2 (ROS2). Mapper 24 can be used to perform node mapping triggered by QoS sensor 23 or ROS2 launcher 26 to deploy ROS2 nodes to edge computing devices and can consider the movement status of mobile vehicles to perform reverse forwarding path planning. Mapper 24 can obtain scheduler parameters through the operation control unit A2 and use these parameters to allocate resources, plan paths, and deploy nodes. Network conditioner 25 can be used to plan flow rules and set flow rules in connector 20. Robot operating system second-generation pusher 26 can receive user commands from the user interface and generate push messages as events to trigger node mapping. Robot operating system second-generation pusher 26 can also import service quality preset conditions from database A1 into service quality sensor 23.

[0019] Connector 20, Robot Operating System 2 (ROS2) node, and Data Distributed Service Exploration and Operation Control Unit A2 reside at the execution layer, while network modeler 21, global network model storage area 22, quality of service sensor 23, mapper 24, network conditioner 25, and Robot Operating System 2 (ROS2) ejector 26 reside at the remote management layer. The operation control unit A2 can be an independent domain that manages all computing resources, including but not limited to cloud computing resources, automated robot factory computing resources, remote computing resources, edge computing resources, or combinations thereof. In one embodiment, the deployment architecture 2 can be multiple, each an independent domain, and interconnected via the Internet.

[0020] In one embodiment, the robot node scheduling system 1 of FIG1 includes a processing device 12. The connector 20, network modeler 21, quality of service sensor 23, mapper 24, network conditioner 25, and robot operation system second-generation launcher 26 can all be executed by the same processing device 12. In another embodiment, the robot node scheduling system 1 of FIG1 includes multiple processing devices 12. The connector 20, network modeler 21, quality of service sensor 23, mapper 24, network conditioner 25, and robot operation system second-generation launcher 26 can each be implemented by a processing device 12, or a portion of the connector 20, network modeler 21, quality of service sensor 23, mapper 24, network conditioner 25, and robot operation system second-generation launcher 26 can be executed by one processing device 12, while the remaining portions can be executed by another or more processing devices 12. Furthermore, the global network model storage area 22 can be implemented by the memory 11 of FIG1. In other words, the global network model storage area 22 and the storage medium storing the multiple instructions can be the same memory 11; or, the global network model storage area 22 and the storage medium storing the multiple instructions can be two separate memory 11s.

[0021] Please refer to Figure 3, which is a flowchart illustrating a robot node scheduling method according to an embodiment of the present invention. As shown in Figure 3, the robot node scheduling method includes: Step S101: establishing a network structure; Step S102: generating a push message according to a user instruction; Step S103: acquiring multiple telemetry data from multiple connectors; Step S105: determining whether the multiple telemetry data meet preset conditions; when the determination result of step S105 is "yes", step S107: generating a scheduling trigger command; Step S109: responding to the scheduling trigger command, obtaining scheduling decision parameters from the operation control unit; Step S111: obtaining the position information of the mobile vehicle from one of the multiple connectors, and taking the one of the multiple connectors as the starting point; Step S113: according to the network structure... And scheduling decision parameters, determine another of the plurality of connectors as the first target point; Step S115: According to the network structure, obtain the shortest path from the starting point to the first target point as the first forwarding path; Step S117: Deploy the second-generation node of the robot operation system to the first target edge computing device; Step S119: Transmit the data from the mobile vehicle to the first target point according to the first forwarding path; Step S121: At least according to the plurality of telemetry data, position information, speed information and scheduling decision parameters, execute the scheduling reverse forwarding algorithm to determine the second forwarding path; and Step S123: Transmit the data from the first target point to the target point indicated by the second forwarding path according to the second forwarding path.

[0022] The condition that the judgment result of steps S102 and S105 is "yes" can occur selectively. That is, when the telemetry data meets multiple preset conditions, or when a push message is generated according to the user's instruction, the system executing this robot node scheduling method will generate a scheduling trigger command. In addition, steps S111, S113, S115, and S117 are node mappings executed after step S109. That is, the scheduling trigger command can trigger the acquisition of scheduling decision parameters and node mapping. The following uses the deployment architecture of Figure 2 as an example to illustrate the steps of Figure 3.

[0023] In step S101, the network modeler 21 establishes a network structure suitable for the operating environment, which includes multiple working areas. Please refer to Figure 4, which is a schematic diagram of a network structure according to an embodiment of the present invention. As shown in Figure 4, the network structure W1 includes multiple connectors L1 to L9 corresponding to the multiple working areas R1 to R9. Each of the multiple connectors L1 to L9 can be connected to the network modeler 21, the quality of service sensor 23, and the network conditioner 25 as shown by connector 20 in Figure 2, and each can be connected to one or more edge computing devices. It should be noted that Figure 4 exemplarily shows nine connectors arranged in an array; however, the present invention does not limit the number or arrangement of connectors, nor does it limit the equidistant spacing between adjacent connectors.

[0024] Please also refer to Figures 2 and 3. In step S102, the second-generation pusher 26 of the robot operating system can receive user instructions and generate push messages accordingly. User instructions can instruct the system's software to update and generate node scheduling.

[0025] In step S103, the Quality of Service (QoS) sensor 23 can obtain telemetry data from each connector 20. Each piece of telemetry data may include one or more telemetry sub-data. The telemetry sub-data may include at least one of the following: hop count, network latency, network loss rate, connection capacity (bandwidth), flow capacity, mobile vehicle computing power, edge computing device computing power, connection lifetime, connection distance, timestamp of Precision Timing Protocol, connector configuration information, connector status information, and mobile vehicle movement trajectory.

[0026] In step S105, the quality of service (QoS) sensor 23 determines whether the multiple telemetry data sets meet preset conditions. The preset conditions may include one or more standard parameters corresponding to the one or more telemetry sub-data sets. The standard parameters may include the upper or lower limit of at least one of the aforementioned telemetry sub-data sets. For example, the preset conditions may include, but are not limited to, network latency greater than or equal to the upper limit of network latency, network loss rate greater than or equal to the upper limit of network loss rate, connection capacity greater than the upper limit of connection capacity, flow capacity greater than the upper limit of flow capacity, computing energy required for local computing and scheduling greater than or equal to the upper limit of computing energy required for mobile vehicles, and / or the computing energy required for tasks processed by the edge computing device greater than the upper limit of computing energy required for the edge computing device.

[0027] When the service quality sensor 23 determines that the multiple telemetry data do not meet the preset conditions, the service quality sensor 23 can execute steps S103 and S105 again.

[0028] When the service quality sensor 23 determines that the multiple telemetry data meet preset conditions, the service quality sensor 23 executes step S107. It should be noted that when the telemetry data meets the preset conditions, step S107 is executed by the service quality sensor 23; when a push message is generated according to a user instruction, the push message can serve as a scheduling trigger instruction. The scheduling trigger instruction is used to trigger the node scheduling of the second generation of the robot operation system.

[0029] In step S109, mapper 24 can obtain the scheduling trigger command generated by service quality sensor 23 through network conditioner 25, or obtain the push message as the scheduling trigger command through robot operating system second-generation pusher 26. In response to the scheduling trigger command, mapper 24 obtains scheduling decision parameters from operation control unit A2 and performs node mapping (steps S111-S117). The scheduling decision parameters indicate the first target edge computing device. The scheduling decision parameters may include identification information of the first target edge computing device. For example, the scheduling decision parameters may specify a Kubernetes basic deployment unit (pod) and the edge computing device running the basic deployment unit as the first target edge computing device.

[0030] In step S111, the mapper 24 obtains the location information of the mobile vehicle from one of the plurality of connectors 20 via the network conditioner 25. The location information may include at least one of location information obtained from a Global Positioning System (GPS) and location information obtained from a wireless network. Specifically, the location information of the mobile vehicle may be recorded in the connector closest to the mobile vehicle among the plurality of connectors 20, and the mapper 24 may obtain the location information of the mobile vehicle from that connector. Furthermore, the mapper 24 sets that connector as a starting point. The starting point can be considered as a data source in the network structure.

[0031] In step S113, the mapper 24 determines one of the plurality of connectors 20 as the first target point based on the network structure and scheduling decision parameters. The first target point corresponds to the first target edge computing device, and the connector of the first target point can be the one closest to the first target edge computing device. The first target point can be regarded as the data destination.

[0032] In step S115, mapper 24 obtains the shortest path from the starting point to the first destination point based on the network structure W1, and uses it as the first forwarding path. The network structure can indicate the connection relationship between the connector 20 at the starting point and the connector 20 at the first destination point. Mapper 24 determines the shortest path from the starting point to the first destination point as the first forwarding path based on the connection relationship. For example, mapper 24 can use, but is not limited to, the Dijkstra algorithm to perform step S115.

[0033] In step S117, mapper 24 deploys the second-generation node of the robot operation system to the first target edge computing device.

[0034] In step S119, the connector 20 at the starting point transmits data from the mobile vehicle along the first forwarding path to the connector 20 at the first target point, enabling the first target edge computing device corresponding to the first target point to receive data from the mobile vehicle and perform calculations. In one embodiment, step S119 may include a network conditioner 25 planning flow rules and placing the flow rules among the plurality of connectors 20 at multiple path points indicated by the first forwarding path. In other words, each connector 20 corresponding to a path point on the first forwarding path may store flow rules. The flow rules include a matching field, an instruction set, and flow rule associations, wherein the matching field indicates the configuration between the plurality of path points, the instruction set indicates the actions of the plurality of path points, and the flow rule association indicates the flow direction of the first forwarding path. Accordingly, the connectors 20 at the plurality of path points on the first forwarding path can forward data according to the flow rules.

[0035] In step S121, the mapper 24 executes a scheduling reverse forwarding algorithm based at least on telemetry data, the location information of the mobile vehicle, the speed information of the mobile vehicle, and scheduling decision parameters to determine the second forwarding path. In one embodiment, the mapper 24 executes the scheduling reverse forwarding algorithm based on telemetry data (including hop count, network latency, network loss rate, connection capacity, flow capacity, mobile vehicle computing power, and edge computing device computing power), the location information of the mobile vehicle, the speed information of the mobile vehicle, the mobile vehicle number, the direction of movement of the mobile vehicle, and scheduling decision parameters to determine the second forwarding path. The scheduling reverse forwarding algorithm is based on a Markov decision process (MDP). In other words, the mapper 24 can predict the location information of the mobile vehicle when the data is delivered to the mobile vehicle based on telemetry data, the location information of the mobile vehicle, the speed information of the mobile vehicle, and scheduling decision parameters, and take the connector closest to the location information as the target point (i.e., the data destination) indicated by the second forwarding path.

[0036] In step S123, the connector 20 transmits data from the first target point to the target point indicated by the second forwarding path according to the second forwarding path, so that the edge computing device corresponding to the target point can receive data from the first target point and perform calculations.

[0037] In one embodiment, before steps S111 and S121, the network regulator 25 can obtain an incoming packet from the connector 20. The incoming packet may contain parameter values ​​indicating the scheduling forwarding calculation in steps S111 to S115 or the scheduling reverse forwarding calculation in step S121. The network regulator 25 can then construct the mapper 24 based on these parameters.

[0038] In summary, the robot node scheduling method and system according to one or more embodiments of the present invention predicts the position of the mobile vehicle and then considers the data transmission delay from the mobile vehicle to the edge computing device and the data transmission delay from the edge computing device to the mobile vehicle. Therefore, the robot node scheduling method and system according to one or more embodiments of the present invention can manage the connectors in the network structure in a flexible and simplified manner, thereby reducing the latency of computing tasks in the robot Internet and improving the overall network performance.

[0039] In one embodiment, step S123 of FIG3 may include network conditioner 25 planning flow rules and placing flow rules in multiple path points indicated by the second forwarding path among the plurality of connectors 20. In other words, each connector 20 corresponding to a path point on the second forwarding path may store flow rules. Flow rules include a matching field, an instruction set, and flow rule associations, wherein the matching field indicates the configuration between the plurality of path points, the instruction set indicates the actions of the plurality of path points, and the flow rule association indicates the flow direction of the second forwarding path. Taking FIG4 as an example, assuming the second forwarding path is from connector L9 through connectors L6 and L5 to connector L4, the path points indicated by the second forwarding path may include connectors L9, L6, L5, and L4. Accordingly, connectors L9, L6, L5, and L4 at multiple path points on the second forwarding path can forward data according to the flow rules.

[0040] Furthermore, at least one of the data from the mobile vehicle and the data from the first target point is distributed data service data, and the network structure is a virtual global area network structure with a location-aware connector. The location-aware connector is used to obtain location information based on at least one of location information from the Global Positioning System and location information obtained from a wireless network.

[0041] Please refer to Figures 2, 4, and 5 together, where Figure 5 is a flowchart illustrating the establishment of a network structure in a robot node scheduling method according to an embodiment of the present invention. As shown in Figure 5, establishing the network structure includes: step S201: obtaining a working environment map and dividing the working environment map into multiple working areas; step S203: configuring the multiple connectors in the multiple working areas to generate first information; step S205: binding multiple edge computing devices to be configured to the multiple connectors to generate second information; step S207: setting weights associated with the multiple connectors according to the multiple telemetry data to generate third information; and step S209: merging the first information, second information, and third information into a global network model. The steps shown in Figure 5 can be executed before step S103 in Figure 3 and can be regarded as a detailed flowchart of an embodiment of step S101 in Figure 3.

[0042] Steps S201-S205 can be performed by the network modeler 21 obtaining the working environment map from the database A1 and executing it accordingly. In step S207, the network modeler 21 can set weights associated with the connectors 20 based on the multiple telemetry data to generate third information. For example, the network modeler 21 can set the weights of each connector 20 and the weights of the connections between two adjacent connectors 20 based on one or more of the network latency rate, network loss rate, connection capacity, and flow capacity of the telemetry data to generate third information.

[0043] In step S209, the network modeler 21 can merge the first information, the second information, and the third information into the global network model storage area 22. Furthermore, the merged global network model storage area 22 can be updated to the database A1. Accordingly, whenever the robot node scheduling system is started, the global network model stored in the database A1 can be imported into the global network model storage area 22.

[0044] Please refer to Figures 2 and 6 together, where Figure 6 is a flowchart illustrating the robot node scheduling method for obtaining a first forwarding path according to an embodiment of the present invention. As shown in Figure 6, obtaining the first forwarding path includes: step S301: extracting a portion of the network structure based on the starting point and the first target point; and step S303: obtaining the shortest path from the starting point to the first target point from the portion of the network structure, as the first forwarding path. The steps shown in Figure 6 can be executed between steps S113 and S117 in Figure 3, and can be regarded as a detailed flowchart of an embodiment of step S115 in Figure 3.

[0045] In step S301, mapper 24 extracts a portion of the network structure containing the starting point and the first target point as a partial network structure. Taking Figure 4 as an example, in one embodiment, assuming the starting point is connector L1 and the first target point is connector L5, the partial network structure can be a structure composed of connectors L1, L2, L4, and L5. In another embodiment, the partial network structure can be a structure composed of connectors L1, L2, L3, L4, L5, and L6. In other words, the partial network structure may include the connectors at the starting point, the connectors at the first target point, and a minimum number of connectors connecting the starting point and the first target point; or, the partial network structure may include a minimum number of connectors and one or more connectors extending outwards.

[0046] In step S303, mapper 24 obtains the shortest path from the starting point to the first target point from a portion of the network structure, and uses it as the first forwarding path. For example, mapper 24 can use the path of connector L1, connector L2 and connector L5 as the first forwarding path; or, mapper 24 can use the path of connector L1, connector L4 and connector L5 as the first forwarding path.

[0047] The robot node scheduling system and method proposed in one or more embodiments of the present invention can realize second-generation node scheduling of robot operation systems with distributed data services. Running in the operation control unit environment, it uses a centralized controller to calculate location-aware task scheduling decisions, wherein the operation control unit can be a K8 cloud-edge intelligent control system. The robot node scheduling system and method proposed in one or more embodiments of the present invention can utilize distributed data service exploration technology in a distributed data service global network, replacing the broadcast method of the second-generation robot operation system, and using a point-to-point pairing method for data transmission, thereby reducing network congestion and improving point-to-point matching efficiency.

[0048] In summary, the robot node scheduling method and system according to one or more embodiments of the present invention provide scheduling for second-generation nodes of robot operation systems in a robot internet under the background of mobile edge computing. Based on network topology and the location information of mobile vehicles, it performs data transmission routing planning, not only planning the scheduling from mobile vehicles to edge computing devices, but also considering the movement status of mobile vehicles during the data transmission process from edge computing devices to mobile vehicles. Therefore, the robot node scheduling method and system according to one or more embodiments of the present invention can manage the connectors in the network structure in a flexible and simplified manner, thereby reducing the latency of computing tasks in the robot internet and improving the overall network performance.

[0049] While the present invention has been disclosed above with reference to the foregoing embodiments, it is not intended to limit the invention. Any modifications and refinements made without departing from the spirit and scope of the invention are within the scope of patent protection of the present invention. For details regarding the scope of protection defined in the present invention, please refer to the appended claims.

[0050] 1: Robot Node Scheduling System 11: Memory 12: Processing device 2: Deployment Architecture 20: Connector 21: Network Modeler 22: Global Network Model Storage Area 23: Service Quality Sensor 24: Mapper 25: Network Conditioner 26: Second-generation pusher for robot operation system A1: Database A2: Operation Control Unit W1: Network Structure R1-R9: Work Area L1-L9: Connectors S101, S102, S103, S105, S107, S109, S111, S113, S115, S117, S119, S121, S123, S201, S203, S205, S207, S209, S301, S303: Steps

Claims

1. A robot node scheduling method, comprising executing the following steps at least once: establishing a network structure, wherein the network structure includes multiple connectors corresponding to multiple work areas; acquiring multiple telemetry data from the connectors and determining whether the telemetry data meets preset conditions; generating a scheduling trigger command when the telemetry data meets the preset conditions or when a push message is generated according to a user instruction; in response to the scheduling trigger command, acquiring a scheduling decision parameter from an operation control unit and executing a node mapping, wherein the scheduling decision parameter indicates a first target edge computing device, and the node mapping includes: acquiring position information of a mobile vehicle from one of the connectors and using that connector as a starting point; determining another connector as a first target point according to the network structure and the scheduling decision parameter, wherein the first target point corresponds to the first target edge computing device; acquiring a shortest path from the starting point to the first target point according to the network structure, as a first forwarding path; and deploying a second-generation Robot Operating System.

2. From the ROS2 node to the first target edge computing device; transmit data from the mobile vehicle to the first target point according to the first forwarding path; execute a scheduling reverse forwarding algorithm to determine a second forwarding path based at least on the telemetry data, the location information, the speed information of the mobile vehicle and the scheduling decision parameters, wherein the scheduling reverse forwarding algorithm is based on a Markov decision process; and transmit data from the first target point to the target point indicated by the second forwarding path according to the second forwarding path.

2. The robot node scheduling method as described in claim 1, wherein establishing the network structure comprises: obtaining a working environment map and dividing the working environment map into working areas; configuring the connectors in the working areas to generate first information; binding multiple edge computing devices to be configured to the connectors to generate second information; setting weights associated with the connectors based on the telemetry data to generate third information; and merging the first information, the second information, and the third information into a global network model.

3. The robot node scheduling method as described in claim 1, wherein each of the telemetry data includes multiple telemetry sub-data, and the preset conditions include multiple standard parameters corresponding to the telemetry sub-data respectively.

4. The robot node scheduling method as described in claim 1, wherein each of the telemetry data includes at least one of network latency, network loss rate, connection capacity, flow capacity, mobile vehicle computing power, and edge computing device computing power.

5. The robot node scheduling method as claimed in claim 1, wherein obtaining the shortest path from the starting point to the first target point as the first forwarding path according to the network structure comprises: extracting a portion of the network structure according to the starting point and the first target point; and obtaining the shortest path from the starting point to the first target point from the portion of the network structure as the first forwarding path.

6. The robot node scheduling method as claimed in claim 1, wherein transmitting the data from the first target point according to the second forwarding path comprises: planning flow rules and placing the flow rules in the connectors at a plurality of path points indicated by the second forwarding path, wherein the flow rules include a matching field, an instruction set, and flow rule associations, wherein the matching field indicates the configuration between the path points, the instruction set indicates the actions of the path points, and the flow rule associations indicate the flow direction of the second forwarding path.

7. The robot node scheduling method as claimed in claim 1, wherein at least one of the data from the mobile vehicle and the data from the first target point is data distribution service (DDS) data, and the network structure is a virtual global network structure with location-aware connectors.

8. A robot node scheduling system, comprising: at least one memory storing a plurality of instructions; and at least one processing device connected to the at least one memory for executing the instructions to perform the robot node scheduling method as described in any one of claims 1-7.

Images