Heterogeneous robot cluster management platform and management method thereof

Abstract

The invention provides a heterogeneous robot cluster management platform and a management method thereof, a scheduling management platform for uniformly managing and scheduling heterogeneous robots is designed, and an MES interface management layer performs data interaction with a factory MES system; the resource management layer manages robot states, task queues, operation resources, data service functions and the like; the scheduling algorithm layer manages and schedules tasks and robot operation according to resource occupation conditions; the task execution layer issues a scheduling instruction to the robot and collects data of the robot and the sensor. The communication management layer realizes identification (such as a known protocol and an unknown protocol), calling and adaptive learning with a robot communication protocol, and data interaction with the user interface layer; the user interface layer presents user-level interface data and imports a data protocol analysis result to the communication management layer. Various robots are incorporated into a unified management scheduling platform, so that the complexity of the system is effectively reduced, and the safety and reliability of the production system are improved.

Description

Technical Field

[0001] The present application relates to the technical field of industrial robot management, and in particular to a heterogeneous robot cluster management platform and a management method thereof. Background Art

[0002] With the continuous advancement of science and technology, in industrial application scenarios, in order to adapt to the automation, flexibility and intelligence of production lines, the scenes often use the production method of integrated operation of automated production lines and industrial robots, especially mobile robots.

[0003] However, in the face of complex automated production systems, robots of different brands and types often work simultaneously in the same scene. In general, robots of different brands have their own scheduling systems to perform on-site management and scheduling of robots of their own brands. This may result in the existence of multiple robot scheduling systems in the same production workshop or scene. The robot scheduling systems are incompatible with each other. The primary feature of the incompatibility is that the communication between the robot and the scheduling system is incompatible, the communication protocol does not adopt the standard communication protocol, and some protocols are private or undisclosed protocols, which increases the complexity of protocol parsing; in addition, if multiple scheduling systems are used, the on-site data interface will be complex and the data node types will be numerous, causing great risks to the reliability of the overall production system and operational safety.

[0004] Based on this, a new heterogeneous robot cluster management platform solution is needed. Summary of the invention

[0005] In view of this, the embodiments of this specification provide a heterogeneous robot cluster management platform and a management method thereof.

[0006] The embodiments of this specification provide the following technical solutions: The embodiment of this specification provides a heterogeneous robot cluster management platform, including: A scheduling management platform for unified management and scheduling of heterogeneous robots of different brands and types; The dispatch management platform includes the following modules: The MES interface management layer is used to exchange data with the factory MES system, receive order instructions and upload operation information. MES stands for Manufacturing Execution System. The resource management layer is used to manage robot status, task queues, operational resources, and provide data analysis services through large language models; The scheduling algorithm layer is used to perform task allocation, route planning, and abnormal scheduling based on task orders and robot resource occupancy; The task execution layer is used to send the instructions generated by the scheduling algorithm layer to the robot, and also collects data from the robot and related sensors; The communication management layer is responsible for identifying and calling the communication interface protocol with the robot; for different communication protocols, it realizes adaptive learning of the communication interface protocol, and realizes the data interaction protocol with the user interface layer to ensure data interaction and scheduling management with the robot; the different communication protocols include known protocols and unknown protocols; The user interface layer is used to present user-level interface data and import relevant data protocol parsing results into the communication management layer.

[0007] The embodiment of this specification also provides a heterogeneous robot cluster management method, including the following steps: Identify the communication protocols of heterogeneous robots through the communication management layer and determine whether they are known protocols or unknown protocols; For unknown protocols, the protocol learning and feature extraction modules are used to learn and extract features, and a model adapted to the protocol is established; Based on the task allocation and route planning of the scheduling algorithm layer, the instructions are sent to the robot through the task execution layer, and the robot data is collected; At the resource management level, a data analysis engine driven by a large language model is used to build an operational knowledge base and generate operational reports and optimization suggestions; The data and analysis results are displayed through the user interface layer, and data is exchanged with the factory MES system through the MES interface management layer.

[0008] Compared with the prior art, the beneficial effects that can be achieved by at least one of the above technical solutions adopted in the embodiments of this specification include at least: 1. This application designs a platform technology that can uniformly manage clusters of heterogeneous robots. It can realize autonomous communication protocol identification and successfully establish communication for robots of different brands, especially robots equipped with different communication protocols, thereby clearing obstacles for inclusion in this system management platform.

[0009] 2. This application uses large model access technology to realize the analysis service function of the management platform for robot operation and production, and provides operation optimization suggestions and analysis reports in the form of natural language, thereby providing factory production managers with operation decision support capabilities.

[0010] 3. This application implements a scheduling and management platform for heterogeneous robot clusters, realizes robot cluster management capabilities, reduces the complexity of intelligent production workshop system equipment management, and improves production reliability and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0012] Figure 1 It is a schematic diagram of the unified cluster scheduling architecture of heterogeneous robots in this application; Figure 2 It is a framework diagram of the dynamic protocol parsing and learning algorithm in this application; Figure 3 It is a diagram of the protocol learning and feature extraction framework in this application; Figure 4 This is a schematic diagram of the management platform combined with heterogeneous cluster robot communication in this application. DETAILED DESCRIPTION

[0013] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0014] The following describes the implementation methods of the present application through specific specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The present application can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that, in the absence of conflict, the features in the following embodiments and embodiments can be combined with each other. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work belong to the scope of protection of the present application.

[0015] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present application, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspect described herein can be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this device and / or practice this method.

[0016] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present application. The illustrations only show the components related to the present application rather than being drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed at will, and the component layout may also be more complicated.

[0017] In industrial application scenarios, in order to adapt to the automation, flexibility and intelligence of production lines, the scenarios often use a production method that integrates automated production lines and industrial robots, especially mobile robots.

[0018] However, in the face of complex automated production systems, robots of different brands and types often work simultaneously in the same scene. In general, robots of different brands have their own scheduling systems to perform on-site management and scheduling of robots of their own brands. However, this may result in the existence of multiple robot scheduling systems in the same production workshop or scene. The robot scheduling systems are incompatible with each other. The primary feature of the incompatibility is that the communication between the robot and the scheduling system is incompatible, the communication protocol does not adopt the standard communication protocol, and some protocols are private or undisclosed protocols, which increases the complexity of protocol parsing; in addition, if multiple scheduling systems are used, the on-site data interface will be complex and the data node types will be numerous, causing great risks to the reliability and operation safety of the overall production system.

[0019] Therefore, a scheduling system platform that is compatible with robots of all brands and compatible with known communication protocols is needed, which can also dynamically parse and identify unknown communication protocols, so as to achieve unified scheduling of robots of all types, greatly reduce the complexity of the overall system, and improve the safety and reliability of the production system.

[0020] Based on this, the embodiments of this specification propose a platform for unified management of clusters of heterogeneous robots. The overall idea is: first, a cluster robot scheduling technology architecture is proposed, which can realize unified scheduling and management of robots of different brands and types; secondly, a dynamic communication protocol parsing and learning algorithm framework is developed, which can automatically identify and adapt the communication protocols of different robots, including dynamic parsing and learning of unknown protocols; finally, a dynamic protocol learning and feature extraction method is designed, which can quickly construct a communication protocol model adapted to heterogeneous robots by separating personalized and general features, thereby incorporating various types of robots into a unified management and scheduling platform, effectively reducing the complexity of the system and improving the safety and reliability of the production system.

[0021] The technical solutions provided by various embodiments of the present application are described below in conjunction with the accompanying drawings.

[0022] like Figure 1As shown, the embodiment of this specification provides a heterogeneous robot cluster management platform and designs a scheduling management platform, which is used to uniformly manage and schedule heterogeneous robots of different brands and types. The scheduling management platform includes an MES interface management layer, a resource management layer, a scheduling algorithm layer, a task execution layer, a communication management layer, and a user interface layer.

[0023] The MES interface management layer is used to interact with the factory MES system, receive order instructions and upload operation information; the resource management layer is used to manage the robot status, task queues, operation resources, and provide data analysis services through a large language model; the scheduling algorithm layer is used to perform task allocation, route planning and exception scheduling according to task orders and robot resource occupancy; the task execution layer is used to send instructions generated by the scheduling algorithm layer to the robot, and also collect data from the robot and related sensor data; the communication management layer is used to identify and call the communication interface protocol between the robot; for different communication protocols, adaptive learning of the communication interface protocol is implemented, as well as data interaction protocols with the user interface layer to ensure data interaction and scheduling management with the robot; different communication protocols include known protocols and unknown protocols; the user interface layer is used to present user-level interface data and import relevant data protocol parsing results into the communication management layer.

[0024] Specifically, the MES interface management layer: the data interaction layer between the cluster scheduling platform and the factory MES system, which is used to receive order and other instruction information from the MES system, and also used to upload robot operation information to the MES data platform. It includes the communication section with MES, the interactive database and cache management section, the production task order analysis section, and the key operation data extraction section.

[0025] Resource management layer: unified management of each robot's status, task order, job type, driving route, endurance, health status, etc.; as well as task queue management, comprehensive operation resource management, and data service function management driven by large language models.

[0026] Scheduling algorithm layer: Based on task orders and robot resource occupancy, overall management and scheduling of task orders and robot operations; the main contents include task balancing, route planning, task order allocation, abnormal situation scheduling and command, etc. It is implemented based on the scheduling algorithm or strategy embedded in the system; it can also be implemented based on a certain learning model.

[0027] Task execution layer: The task execution layer mainly sends the instructions generated by the scheduling algorithm layer to the robot, and also collects data from the robot and related sensors.

[0028] Communication management layer: includes the identification and calling of the communication interface protocol with the robot, the adaptive learning of the interface protocol, and the implementation of the data interaction protocol with the user interface layer.

[0029] User interface layer: implements the presentation of user-level interface data and imports relevant data protocol parsing results into the communication management layer.

[0030] In some embodiments, the resource management layer introduces a data analysis engine driven by a large language model to build an operation knowledge base, wherein data such as fault logs, operation trajectories, production orders, etc. are used to build the operation knowledge base; multimodal data analysis is performed based on the large language model to generate operation reports, optimization suggestions and natural language reports to provide decision support for production managers.

[0031] like Figure 1 As shown in the figure, at the resource management level, a data analysis engine driven by a large language model is introduced. The main purpose is to provide users with operational service functions, such as operational report generation and optimization suggestions.

[0032] Driven by the large language model data analysis engine, an operation knowledge base is constructed based on the operation data of various heterogeneous robots and data from the MES system. The main contents of the operation knowledge base include: the digitized system includes fault logs of each robot and production line, trajectory data of each robot's operation route, production order data from the MES system and production task order data of each robot, and production data of the production line such as beat, status, load and other productive databases.

[0033] Based on the operation knowledge base, a large model is used for multimodal analysis to analyze and integrate data of various time, space and point dimensions, and finally obtain analysis results, including overall statistics of operation data, order relationships, generation of reports, etc.; as well as reports on operation optimization, etc.

[0034] In some embodiments, the communication management layer includes: a protocol discovery and identification layer, which is used to discover a communication protocol different from the embedded standard communication protocol and identify the protocol; when it is identified as a known protocol, communication with the robot is directly achieved through the communication layer; the protocol learning and feature extraction layer, which is used to establish a protocol model in the protocol model construction layer when it is identified as an unknown protocol, and adapt and understand the unknown protocol by building a successful protocol model; a protocol adaptation and conversion layer, which is used to convert task instructions into a protocol format used by robots equipped with different communication protocols; a data stream parsing and conversion layer, which is used to parse the data stream information fed back by the robot and convert it into a unified format compatible with the platform; a fault tolerance and recovery layer, which is used to handle errors in the communication process and ensure the integrity and accuracy of the communication data; a communication layer, which is used to communicate with each robot and perform data transmission tasks; a user interface layer, which is used to present user-level interface data and import relevant data protocol parsing results into the communication management layer.

[0035] In this application, one of the core purposes is to realize data interaction among heterogeneous robots based on the scheduling platform, so as to ultimately realize a unified scheduling function.

[0036] However, different robots come from different brands and perform different types of tasks. The communication protocols integrated into each robot may not be the same, or even compatible with standard communication protocols. To use the same scheduling platform to achieve unified scheduling of all these robots, it is first necessary to achieve successful communication between the scheduling platform and each robot in order to exchange data.

[0037] To achieve the above objectives, the scheduling platform in this application is based on the user interface layer and has an embedded protocol discovery and identification layer.

[0038] like Figure 2 As shown, the main function of the protocol discovery and identification layer is to discover the existence of a communication protocol that is different from the embedded standard communication protocol (known protocol), identify the protocol, and successfully adapt the protocol through the learning model, thereby establishing the normal communication function between the platform and the robot device, and ultimately achieving the purpose of data interaction and scheduling management of the robot.

[0039] like Figure 2 As shown, if it is identified as a known protocol (the scheduling platform has embedded standard communication protocols), communication with the robot is achieved directly through the communication layer.

[0040] If an unknown protocol is identified, a protocol model is established in the protocol model construction layer based on the results of the protocol learning and feature extraction layer. Successful construction of the protocol model means that the basic conditions for realizing the communication function have been met. After subsequent processing by the adaptation and conversion layer, data stream parsing and conversion layer, and fault tolerance and recovery layer, communication can be achieved.

[0041] Regarding the protocol learning and feature extraction layer, the scheduling platform attempts to communicate with the robot of the corresponding unknown protocol. The robot reports the protocol type by itself, uses the built-in protocol detection algorithm to extract the feature text or feature data format of the protocol, and continuously modifies the protocol model, and finally establishes a protocol format or protocol model that is suitable for the robot.

[0042] in Figure 2 In the protocol adaptation and conversion layer, this layer converts the task instructions of the scheduling platform into the protocol format used by robots equipped with different communication protocols to adapt to the protocols used by different robots.

[0043] Data stream parsing and conversion layer: This layer is used to parse the data stream information fed back by the robot and convert it into a unified format compatible with the platform.

[0044] Fault tolerance and recovery layer: This layer handles errors in the protocol parsing process, such as data loss, communication interruption, etc., to ensure the integrity and accuracy of communication data.

[0045] The communication layer is used to communicate with each robot and perform data transmission tasks.

[0046] In some embodiments, the protocol learning and feature extraction layer adopts a personalized and general feature separation method, including: general feature extraction, which is used to identify structured information in data packets, wherein the structured information is general features, including: field length, repetition pattern, and flag bit; personalized protocol learning, which is used to learn and extract personalized features of the communication protocol based on capturing the message data stream sent back by the robot; wherein the personalized features include the number, type, and format information of the fields of the personalized protocol; personalized feature extraction and protocol adaptation, which are used to generate parsing rules for protocol fields, wherein the parsing rules for protocol fields include field offset, data type, and verification rules; protocol encapsulation, which is used to combine general features and personalized features to build a complete communication protocol model to achieve communication adaptation with heterogeneous robots.

[0047] In this application, the scheduling platform connects with the robot by trying to communicate and captures the robot's return message data stream. The method for constructing the unknown protocol model is to use a personalized-feature separation method to divide the unknown protocol feature extraction process into general feature extraction and personalized feature learning and extraction.

[0048] Among them, general feature extraction identifies structured information in data packets, such as field length, repetitive pattern, flag bit, etc. CNN or Transformer (attention mechanism) can be used to extract common features of the protocol.

[0049] Personalized protocol learning is to learn the content of the communication protocol in addition to the general features. Based on more captured message data streams, the personalized features are learned and the number, type, and format information of the personalized protocol fields are extracted.

[0050] Specifically, personalized protocol learning adopts meta-learning training method, that is, Figure 3 The “meta-learning training and fast adaptation” shown here means using a meta-learning method to train a protocol parsing model with strong generalization ability, so as to achieve the purpose of more quickly parsing the personalized protocol structure of protocol data based on a small number of samples.

[0051] Personalized feature extraction and protocol adaptation are used to quickly generate parsing rules for protocol fields, such as field offsets, data types, verification rules, etc.

[0052] When the general feature extraction and personalized protocol feature extraction are finally completed, the two are combined and the protocol is encapsulated to form a complete communication protocol that can communicate with the original unknown protocol robot, thereby incorporating the heterogeneous robot into the management and scheduling platform for unified control, achieving the purpose of a management platform to manage and schedule various types of heterogeneous robots.

[0053] In some embodiments, the personalized protocol learning adopts a meta-learning method, and the protocol parsing model is trained through a MAML / Reptile learning algorithm to quickly parse the personalized features of the protocol based on a small number of samples.

[0054] This application adopts the meta-learning method in the unified management and scheduling scenarios of heterogeneous robots, and only requires a small amount of sample data to train the protocol model, so that it can quickly learn and parse the personalized features of new unknown protocols.

[0055] like Figure 3 As shown, for the protocol parsing scenario of this application, each task is for a specific communication protocol, and the specific steps are as follows: Packet preprocessing: Process the raw data stream, decompose the data packets into a format that can be used for feature extraction and personalized processing, and extract metadata (such as protocol identifier, packet length, etc.).

[0056] A neural network based on LSTM or attention mechanism selects a model suitable for protocol parsing. Through meta-learning training and fast adaptation, MAML or Reptile algorithm is used for meta-learning training to optimize the initial parameters of the model so that it can quickly adapt to new protocols with a small number of samples. In the inner loop, the support set is used to perform gradient updates on the model to obtain task-specific parameters. In the outer loop, the query set is used to evaluate the model performance and update the initial parameters.

[0057] Common feature extraction: Extract common features in protocol data, such as bit alignment information, field ordering, and other common structures in data packets.

[0058] Personalized learning module, using LSTM / attention mechanism to analyze protocol change patterns.

[0059] Personalized protocol learning, using meta-learning training, such as Figure 3 The meta-learning training and rapid adaptation shown in the figure uses the MAML / Reptile learning algorithm to train the protocol model, so that the protocol can be quickly learned and parsed based on a small number of samples. When facing a new unknown protocol, a small amount of sample data is used to quickly fine-tune the model so that it can quickly parse the personalized features of the protocol.

[0060] Personalized feature extraction and protocol adaptation to identify key personalized fields and behavioral pattern analysis of the protocol.

[0061] In some embodiments, the scheduling algorithm layer implements task balancing, route planning, task allocation and abnormal scheduling command based on the scheduling algorithm or learning model embedded in the system.

[0062] Specifically, the scheduling algorithm layer is implemented based on the scheduling algorithm or strategy embedded in the system; it can also be implemented based on a certain learning model. The main implementation contents include task balancing, route planning, task order allocation, abnormal situation scheduling command, etc.

[0063] In some embodiments, the fault tolerance and recovery layer includes a heterogeneous communication fault tolerance and recovery layer and a standard communication fault tolerance and recovery layer. The heterogeneous communication fault tolerance and recovery layer is used to specifically handle errors and recovery in heterogeneous robot communications and runs in parallel with the standard communication fault tolerance layer; the standard communication fault tolerance and recovery layer is used to handle errors and recovery of standard communication protocols in communication.

[0064] like Figure 4 As shown, in this application, the communication layer of the scheduling management platform is the execution layer of each communication protocol and the data interface with the scheduling management platform. The communication layer communicates with each robot.

[0065] The heterogeneous communication fault tolerance and recovery layer is specifically designed for the communication of heterogeneous robots. It is parallel to the standard communication fault tolerance and recovery layer and does not affect each other.

[0066] In combination with the above embodiments, the present application also provides a heterogeneous robot cluster management method, which specifically includes the following steps: identifying the communication protocol of the heterogeneous robot through the communication management layer, and judging whether it is a known protocol or an unknown protocol; for unknown protocols, learning and feature extraction are performed through the protocol learning and feature extraction module, and a model adapted to the protocol is established; based on the task allocation and route planning of the scheduling algorithm layer, instructions are sent to the robot through the task execution layer, and robot data is collected; at the resource management layer, a data analysis engine driven by a large language model is used to build an operation knowledge base and generate operation reports and optimization suggestions; data and analysis results are displayed through the user interface layer, and data is interacted with the factory MES system through the MES interface management layer.

[0067] In some embodiments, the protocol learning and feature extraction steps adopt a personalized and universal feature separation method, including universal feature extraction and personalized protocol learning.

[0068] In combination with the above embodiments, the present application adopts a personalized and universal feature separation method, and the protocol learning and feature extraction steps can efficiently process unknown protocols. Universal feature extraction provides basic structural information of the protocol, while personalized protocol learning captures the uniqueness of the protocol. Combining these two features, a protocol parsing model that can communicate with heterogeneous robots can be quickly constructed, thereby realizing unified scheduling and management of heterogeneous robots.

[0069] In some embodiments, during the communication process, communication errors are handled by a fault tolerance and recovery layer to ensure the integrity and accuracy of the data.

[0070] like Figure 4 As shown, in the present application, the fault tolerance and recovery layer includes a heterogeneous communication fault tolerance and recovery layer and a standard communication fault tolerance and recovery layer.

[0071] The communication layer of the dispatch management platform is the execution layer of each communication protocol and the data interface with the dispatch management platform. The communication layer communicates with each robot.

[0072] The heterogeneous communication fault tolerance and recovery layer is specifically designed for the communication of heterogeneous robots. It is parallel to the standard communication fault tolerance and recovery layer and does not affect each other.

[0073] The protocol model in this application is a broader model that describes the overall structure and rules of the communication protocol. It includes all the features of the protocol (such as common features and personalized features) and is used to define and manage the protocol for communicating with the robot, ensuring that the scheduling platform can communicate effectively with the robot. The protocol parsing model is a model used to parse protocol data. Its main task is to extract useful information from the received data stream and convert it into a format that the platform can understand and process. The protocol parsing model is part of the protocol model and is used to implement the parsing rules defined in the protocol model. The training and updating of the protocol parsing model can dynamically affect the performance of the protocol model, enabling it to better adapt to new protocol formats.

[0074] This application realizes unified management and scheduling of robots of different brands and types through a cluster robot scheduling technology architecture, simplifying the complexity of multiple robot scheduling systems in a production workshop or scene.

[0075] Among them, the data analysis engine driven by the large language model is used to analyze the robot operation and production, optimize the production process, and improve production efficiency. Operation optimization suggestions and analysis reports are provided in the form of natural language to help factory production managers make more informed operational decisions. Cluster management reduces the complexity of intelligent production workshop system equipment management, thereby improving production reliability and safety.

[0076] The dynamic communication protocol parsing and learning algorithm framework reduces the time for manual testing and debugging, and quickly adapts to the communication protocols of different robots. It is compatible with known communication protocols and can dynamically parse and identify unknown communication protocols, reducing the complexity of communication management. It can identify and adapt to different communication protocols, improve the compatibility and flexibility of the system, and adapt to changing industrial application scenarios.

[0077] Through a unified scheduling platform, the overall system complexity is reduced and equipment management is simplified. The multimodal data analysis layer based on large models can process and analyze diverse data from different robots and MES systems. In particular, it has practical application value in improving the production level of automated production lines and industrial robots, optimizing resource management, and enhancing decision support in industrial application scenarios.

[0078] In this specification, the same or similar parts between the various embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the embodiments described later, the description is relatively simple, and the relevant parts can be referred to the partial description of the previous embodiments.

[0079] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims

1. A heterogeneous robot cluster management platform, characterized in that: include: A scheduling management platform for unified management and scheduling of heterogeneous robots of different brands and types; The dispatch management platform includes the following modules: MES interface management layer, used to interact with the factory MES system, receive order instructions and upload operation information; The resource management layer is used to manage robot status, task queues, operational resources, and provide data analysis services through large language models; The scheduling algorithm layer is used to perform task allocation, route planning, and abnormal scheduling based on task orders and robot resource occupancy; The task execution layer is used to send the instructions generated by the scheduling algorithm layer to the robot, and also collects data from the robot and related sensors; The communication management layer is used to identify and call the communication interface protocol with the robot; for different communication protocols, it realizes adaptive learning of the communication interface protocol and the implementation of the data interaction protocol with the user interface layer to ensure data interaction and scheduling management with the robot; the different communication protocols include known protocols and unknown protocols; The user interface layer is used to present user-level interface data and import relevant data protocol parsing results into the communication management layer.

2. The heterogeneous robot cluster management platform according to claim 1, characterized in that: The communication management layer includes: The protocol discovery and identification layer is used to discover and identify communication protocols that are different from the embedded standard communication protocol; when it is identified as a known protocol, it directly communicates with the robot through the communication layer; The protocol learning and feature extraction layer is used to establish a protocol model in the protocol model building layer when an unknown protocol is identified, and to adapt and understand the unknown protocol by building a successful protocol model; The protocol adaptation and conversion layer is used to convert task instructions into the protocol format used by robots equipped with different communication protocols; The data stream parsing and conversion layer is used to parse the data stream information fed back by the robot and convert it into a unified format compatible with the platform; Fault tolerance and recovery layer, used to handle errors in the communication process and ensure the integrity and accuracy of communication data; The communication layer is used to communicate with each robot and perform data transmission tasks.

3. The heterogeneous robot cluster management platform according to claim 2, characterized in that: The protocol learning and feature extraction layer adopts a personalized and universal feature separation method, including: General feature extraction is used to identify structured information in data packets, where structured information is general features, including: field length, repetition pattern, and flag bit; Personalized protocol learning is used to learn and extract personalized features of the communication protocol based on the message data stream sent back by the captured robot; the personalized features include the number, type, and format information of the fields of the personalized protocol; Personalized feature extraction and protocol adaptation are used to generate parsing rules for protocol fields, where the parsing rules for protocol fields include field offset, data type, and verification rules; Protocol encapsulation is used to combine common features and personalized features to build a complete communication protocol model and achieve communication adaptation with heterogeneous robots.

4. The heterogeneous robot cluster management platform according to claim 3, characterized in that: The personalized protocol learning adopts a meta-learning method, and trains the protocol parsing model through the MAML / Reptile learning algorithm to quickly parse the personalized features of the protocol based on a small number of samples.

5. The heterogeneous robot cluster management platform according to claim 1, characterized in that: The resource management layer introduces a data analysis engine driven by a large language model to build an operation knowledge base, wherein fault logs, operation trajectories, and production order data are used to build the operation knowledge base; Based on the large language model, multimodal data analysis is performed to generate operation reports, order relationships, and operation optimization suggestion reports to provide decision support for production managers.

6. The heterogeneous robot cluster management platform according to claim 1, characterized in that: The scheduling algorithm layer implements task balancing, route planning, task allocation and abnormal scheduling command based on the scheduling algorithm or learning model embedded in the system.

7. The heterogeneous robot cluster management platform according to claim 2, characterized in that: The fault tolerance and recovery layer includes a heterogeneous communication fault tolerance and recovery layer and a standard communication fault tolerance and recovery layer. The heterogeneous communication fault tolerance and recovery layer is used to specifically handle errors and recovery in heterogeneous robot communications and runs in parallel with the standard communication fault tolerance and recovery layer. The standard communication fault tolerance and recovery layer is used to handle errors and recovery of the standard communication protocol during communication.

8. A heterogeneous robot cluster management method, characterized in that: The following steps are involved: Identify the communication protocols of heterogeneous robots through the communication management layer and determine whether they are known protocols or unknown protocols; For unknown protocols, the protocol learning and feature extraction modules are used to learn and extract features, and a model adapted to the protocol is established; Based on the task allocation and route planning of the scheduling algorithm layer, the instructions are sent to the robot through the task execution layer, and the robot data is collected; At the resource management level, a data analysis engine driven by a large language model is used to build an operational knowledge base and generate operational reports and optimization suggestions; The data and analysis results are displayed through the user interface layer, and data interaction is carried out with the factory MES system through the MES interface management layer.

9. The heterogeneous robot cluster management method according to claim 8, characterized in that: The protocol learning and feature extraction steps adopt a personalized and universal feature separation method, including universal feature extraction and personalized protocol learning.

10. The heterogeneous robot cluster management method according to claim 8, characterized in that: During the communication process, communication errors are handled through the fault tolerance and recovery layer to ensure the integrity and accuracy of the data.

Images