Storage device and method for artificial intelligence operation and maintenance robot

Abstract

The invention discloses a storage device and method for an artificial intelligence operation and maintenance robot, which are deployed in a data center or a cloud platform and are used for carrying out centralized operation and maintenance management on target hosts in a plurality of service systems. The system operation and maintenance server comprises a resource registration module, a monitoring acquisition module, a configuration management module, a task arrangement module, an alarm and self-healing module, an operation and maintenance auditing module, an interface management module and an identity authentication and authority control module. The resource registration module is used for receiving information of a target host and a business system to which the target host belongs, generating a unique resource identifier and establishing a resource information base; the monitoring acquisition module issues a monitoring acquisition strategy based on the resource identifier and acquires monitoring index data and an operation log; the configuration management module performs difference comparison on the current configuration and the configuration baseline, and issues a configuration template and performs configuration version management after confirmation; and the task arrangement module arranges the atomic operation and maintenance operation into a task execution plan containing an execution sequence and a rollback strategy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent operation and maintenance, and in particular to a storage device and method for an artificial intelligence operation and maintenance robot. BACKGROUND

[0002] The existing pick-and-place management solutions for operation and maintenance robots or automated warehousing mostly focus on simple inventory registration and instruction issuance, lack unified data modeling of storage locations, items, capacities and availability states, and thus make it difficult to adapt to dynamic changes as the target positioning relies on manual configuration or static table entries. In multi-robot concurrent or task-intensive scenarios, there is often a lack of pre-occupancy / mutual exclusion conflict resolution mechanisms, which can easily result in repeated allocation, incorrect picking or disordered picking and placing of the same item. In terms of permission control, some solutions only perform coarse-grained account verification, lack short-term authorization and policy constraints bound to single pick-and-place requests, and thus it is difficult to achieve fine-grained controlled pick-and-place and responsibility tracing according to tasks and quantities. In terms of state verification, single sensing or only relying on robot feedback is often used, and there is a lack of multi-source fusion and confidence management, which makes it difficult to accurately discover in-place, quantity and abnormalities in a timely manner. In terms of inventory updating, many systems do not transactionalize the pick-and-place process, lack atomic boundaries, rollback mechanisms and idempotent processing, and thus are prone to cause inventory errors, repeated deductions or data inconsistencies when there is communication interruption, retransmission or repeated confirmation, thereby affecting operation reliability and continuous automated operation. SUMMARY

[0003] The present application provides a storage device and method for an artificial intelligence operation and maintenance robot.

[0004] The present application solves the above technical problems by the following technical solutions:

[0005] A storage device for an artificial intelligence operation and maintenance robot, comprising:

[0006] A storage resource management unit for data modeling of at least one storage resource set, the storage resource set comprising at least a plurality of storage location identifiers, and item identifiers, capacity constraints and availability states associated with each storage location identifier;

[0007] A task scheduling unit for receiving a pick-and-place request initiated by an operation and maintenance robot, and generating a pick-and-place task plan based on the storage resource set, the pick-and-place task plan comprising at least a target item identifier, a target storage location identifier, a pick-and-place action sequence and a completion criterion;

[0008] A permission and policy control unit for performing authorization verification on the pick-and-place request, and imposing access policies and security constraints on the pick-and-place task plan;

[0009] The state sensing and verification unit is configured to acquire sensing data corresponding to the target storage location identifier and / or the target item identifier before, during and / or after the picking and placing operation, and output a picking and placing consistency result;

[0010] The transaction and ledger unit is configured to map a picking and placing request to an inventory transaction record, and perform a commit update or a rollback update on the inventory state according to the picking and placing consistency result;

[0011] The communication and collaboration interface unit is configured to communicate with the maintenance robot and / or the maintenance management platform, to issue the picking and placing task plan and picking and placing guidance information to the maintenance robot, and receive execution feedback.

[0012] The task arrangement unit is configured to perform pre-occupation on the target item identifier and / or the target storage location identifier after the authorization verification passes, and release the pre-occupation after the transaction and ledger unit completes the commit update.

[0013] In some embodiments, the storage resource management unit maintains a storage location topology relationship and conflict constraints, and the task arrangement unit determines the target storage location identifier and / or generates the picking and placing action sequence based on the topology relationship and conflict constraints.

[0014] In some embodiments, the task arrangement unit performs two-stage arrangement on the picking and placing request: the first stage generates a candidate target storage location set and performs pre-occupation; the second stage determines the final target storage location identifier and solidifies the picking and placing task plan after receiving the result of the pre-picking and placing verification.

[0015] In some embodiments, the permission and policy control unit is configured to generate a short-term authorization token corresponding to a picking and placing request, the short-term authorization token at least including a request number, a robot identifier, a target item identifier, a target quantity and a validity period; and the communication and collaboration interface unit only allows issuing the picking and placing task plan and / or receiving the picking and placing completion confirmation when the short-term authorization token is valid.

[0016] In some embodiments, the state sensing and verification unit includes a multi-source sensing fusion sub-unit configured to fuse sensing data of at least two different modalities to output a fusion determination result of the in-place state, quantity change and / or abnormal state of the target item.

[0017] In some embodiments, the multi-source sensing fusion sub-unit calculates a confidence level for the fusion determination result, and triggers at least one of re-acquisition, switching to a backup verification path or upgrading to manual review when the confidence level is below a threshold.

[0018] In some embodiments, the transaction and ledger unit is configured to establish an atomic transaction boundary for each pick-and-place request, and record pre-pick inventory snapshot, post-pick inventory snapshot, verification result and exception code within the transaction boundary to support rollback update and audit traceability.

[0019] In some embodiments, the transaction and ledger unit adopts an idempotent commit mechanism, de-duplicates the duplicate pick-and-place completion confirmations based on request number, and only performs inventory state commit update when not committed.

[0020] A storage method based on a storage device, comprising:

[0021] S1, receiving a pick-and-place request sent by an operation and maintenance robot, the pick-and-place request at least containing target item identification and target quantity;

[0022] S2, performing authorization verification on the pick-and-place request, and performing pre-occupation on the target item identification and / or candidate target storage location identification after the verification passes;

[0023] S3, generating a pick-and-place task plan based on a set of storage resources, and issuing the pick-and-place task plan and pick-and-place guidance information to the operation and maintenance robot;

[0024] S4, collecting perception data before, during and / or after pick-and-place execution and outputting pick-and-place consistency results;

[0025] S5, writing the pick-and-place request into inventory transaction records, and performing commit update or rollback update on inventory state according to the pick-and-place consistency results;

[0026] S6, releasing the pre-occupation and generating a ledger record, and returning the results to the operation and maintenance robot and / or operation and maintenance management platform.

[0027] In some embodiments, when at least one of the following exceptions is detected: verification inconsistency, pre-occupation timeout, robot execution failure or communication interruption, an exception handling process is triggered, which at least includes: revoking a short-term authorization token, freezing the target item identification and / or target storage location identification, performing rollback update and generating an alarm event.

[0028] The storage device for the artificial intelligence operation and maintenance robot disclosed in the application abstracts one taking and placing operation into a stock transaction that can be preoccupied, verified, submitted and rolled back through a closed-loop architecture of "storage resource modeling-task scheduling-authorization strategy-state perception verification-transactionized account updating-communication coordination": preoccupation of the target article and / or storage location is performed after authorization verification, and two-stage scheduling is combined, which can effectively reduce the conflicts and wrong taking caused by concurrent taking and placing of multiple robots; based on short-term authorization tokens and policy constraints, controlled taking and placing according to the robot identity, task credentials, time window and quantity threshold are realized, and safety and traceability are improved; through multi-source perception fusion and confidence determination, consistency verification of the in-place state and quantity change before and after taking and placing can be performed, and when the confidence is insufficient, reacquisition or backup verification path is triggered, thereby significantly reducing the risk of missed detection, misjudgment and abnormality undetected; through atomic transaction boundary, the snapshots before and after taking and placing and abnormal codes are recorded, and an idempotent submission mechanism is used, which can avoid repeated deduction of stock and repeated accounting in network retransmission or repeated confirmation scenarios, ensure stock consistency, quickly roll back, freeze the target and alarm when an abnormality occurs, realize stable and reliable intelligent taking and placing and stock management, and improve operation and maintenance efficiency and system robustness. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only a part of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of the provided drawings.

[0030] Figure 1 The system overall architecture schematic diagram of the present application is shown.

[0031] Figure 2 The method flow chart of the present application is shown. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0033] Embodiment one

[0034] Referring to Figure 1 and Figure 2The application provides a storage device for an artificial intelligence operation and maintenance robot, which can be deployed in a computer room, a data center, a factory warehouse or the like, and communicates with the operation and maintenance robot and an operation and maintenance management platform through a wired or wireless network. The storage device comprises a storage resource management unit, a task scheduling unit, a permission and policy control unit, a state sensing and verification unit, a transaction and account unit and a communication and cooperation interface unit. Each unit can be implemented by a processor executing a computer program stored in a memory, or can be implemented in the form of an edge computing node, a server or a controller, and the application does not limit this.

[0035] The storage resource management unit is used for data modeling of at least one storage resource set, and the storage resource set at least comprises a plurality of storage location identifiers, each of which is associated with an item identifier, a capacity constraint and an availability state. For example, a data table or an object model records the following fields: storage location identifier LocID, item identifier ItemID, capacity constraint Cap, availability state Avail and current quantity Qty. The availability state can at least include available, unavailable, frozen, in maintenance, etc. To support task planning, the storage resource management unit can also maintain storage location topology relationships and conflict constraints, such as reachable relationship, parallel conflict relationship and taking and placing cost parameters, for calling by the task scheduling unit to reduce the generation probability of unreachable or conflicting tasks.

[0036] The operation and maintenance robot sends a taking and placing request through the communication and cooperation interface unit, and the taking and placing request at least contains a target item identifier and a target quantity, and can optionally contain a robot identifier, a task voucher, a priority, a time window, etc. The permission and policy control unit performs authorization verification on the taking and placing request, and applies access policies and security constraints to the taking and placing task. The access policies can at least include one or more of robot identity, task voucher, time window, taking and placing quantity threshold. After the authorization verification is passed, the task scheduling unit performs pre-occupation on the target item identifier and / or candidate target storage location identifier, so that the same item or the same location is logically locked within the processing period of the request, thereby reducing repeated allocation, wrong taking or preemption caused by multiple request concurrency.

[0037] To further strengthen the security control, in a preferred implementation, the permission and policy control unit generates a short-term authorization token corresponding to a taking and placing request, and the short-term authorization token at least contains a request number, a robot identifier, a target item identifier, a target quantity and a validity period (optionally containing a verification code or a signature field). The communication and cooperation interface unit only allows the taking and placing task plan to be issued and / or the taking and placing completion confirmation to be received when the short-term authorization token is valid, thereby realizing fine-grained authorization and traceability bound to a single task.

[0038] The task scheduling unit is configured to receive the pick-and-place request and generate a pick-and-place task plan. To improve adaptability to dynamic inventory and state changes, the task scheduling unit can adopt a two-stage scheduling: in the first stage, a set of candidate target storage locations is generated based on the set of storage resources (and topological relationship / constraint conflicts), and pre-occupancy is performed; in the second stage, after the state perception and verification unit outputs a pass of the pre-pick-and-place verification, a final target storage location identifier is determined, and the pick-and-place task plan is solidified. The pick-and-place task plan at least includes a target item identifier, a target storage location identifier, a pick-and-place action sequence, and a completion criterion; wherein the action sequence can be abstracted as the logical steps of “arriving at the target location—performing pick-and-place—reporting execution feedback—waiting for verification—ending”, and the completion criterion can be abstracted as “verification passed and inventory transaction successfully committed”.

[0039] After the task scheduling unit generates the pick-and-place task plan, the communication and coordination interface unit issues the pick-and-place task plan and pick-and-place guidance information to the operation robot. The pick-and-place guidance information can at least include the target storage location identifier and the guidance parameters related to the pick-and-place action (such as one or more of the reference coordinates / target pose, the allowed operation window, or the obstacle avoidance constraint), to support the operation robot to complete the coordinated execution of arrival and pick-and-place actions. The present application does not limit the specific expression format of the guidance information.

[0040] The state perception and verification unit is configured to obtain perception data before, during, and / or after pick-and-place execution, and output a pick-and-place consistency result. The perception data can include but is not limited to: item identity recognition results, in-place state detection results, quantity change results, etc. To improve verification reliability, the state perception and verification unit can include a multi-source perception fusion sub-unit that fuses perception data of at least two different modalities, and outputs a fusion determination result of the in-place state, quantity change, and / or abnormal state of the target item.

[0041] In a preferred implementation, the multi-source perception fusion sub-unit calculates a confidence level for the fusion determination result, and triggers at least one processing strategy when the confidence level is below a threshold: re-acquisition, switching to a backup verification path, or upgrading to manual review. Through the above mechanism, the robustness of pick-and-place verification can be improved by reducing missed detection and misjudgment caused by single perception error.

[0042] The transaction and ledger unit is configured to map each pick-and-place request to an inventory transaction record, and perform commit update or rollback update on the inventory state according to the pick-and-place consistency result. To ensure consistency and traceability, the transaction and ledger unit establishes an atomic transaction boundary for each pick-and-place request, and records information such as pre-pick-and-place inventory snapshot, post-pick-and-place inventory snapshot, verification result, and exception code within the transaction boundary, to support audit traceability and exception rollback.

[0043] In a preferred implementation, the transaction and ledger units adopt an idempotent commit mechanism: for duplicate pick-up and drop-off completion confirmations (e.g., network jitter, retransmission, or robot duplicate reporting), the corresponding transaction is determined based on the request number; if it has been committed, the committed result is returned and the inventory status update is no longer performed, thereby avoiding the accounting error caused by duplicate deductions or duplicate entries.

[0044] Once the transaction is successfully submitted, the task orchestration unit releases the pre-occupancy of the target item identifier and / or the target storage location identifier, and the communication and collaboration interface unit sends the execution result, verification result, and inventory update result back to the operation and maintenance robot and / or operation and maintenance management platform to complete the closed loop.

[0045] When at least one of the following anomalies occurs: verification inconsistency, pre-occupancy timeout, robot execution failure, or communication interruption, the device triggers the anomaly handling procedure. The anomaly handling procedure includes at least: revoking the short-term authorization token (if already generated), freezing the target item identifier and / or target storage location identifier, performing a rollback update by the transaction and ledger unit, and generating an alarm event to be sent back to the operation and maintenance management platform. Through freezing and rollback, risks can be limited to a local resource scope, preventing the anomaly from spreading and ensuring that the inventory status is recoverable.

[0046] Based on the aforementioned storage device, a corresponding storage method can be executed. The process can be summarized as follows: receiving a pick-up / placement request → authorization verification → pre-occupancy → generating and issuing a pick-up / placement task plan and guidance information → collecting and verifying sensing data → writing to inventory transactions and committing or rolling back → releasing pre-occupancy and sending back the results. Among these, short-term authorization tokens, two-phase orchestration, multi-source fusion and confidence triggering, atomic transaction boundaries and idempotent commit, as well as abnormal freezing / rollback / alarm mechanisms can work together in the same process, thereby achieving a secure, controllable, verifiable, traceable, and consistent inventory management closed loop for pick-up / placement by maintenance robots.

[0047] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0048] The above description of the disclosed embodiments enables those skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A storage device for an artificial intelligence operation and maintenance robot, characterized in that, include: A storage resource management unit is used to perform data modeling on at least one set of storage resources, wherein the set of storage resources includes at least multiple storage location identifiers, as well as item identifiers, capacity constraints, and availability status associated with each storage location identifier; The task orchestration unit is used to receive the pick-up and place-up request initiated by the operation and maintenance robot, and generate a pick-up and place-up task plan based on the storage resource set. The pick-up and place-up task plan includes at least the target item identifier, the target storage location identifier, the pick-up and place-up action sequence and the completion criteria. The permission and policy control unit is used to perform authorization verification on the access and release request, and to apply access policies and security constraints to the access and release task plan; The status perception and verification unit is used to acquire perception data corresponding to the target storage location identifier and / or the target item identifier before, during and / or after the retrieval and placement are performed, and output the retrieval and placement consistency result. The transaction and ledger unit is used to map a single pick-up and drop-off request to an inventory transaction record, and to perform a commit update or rollback update on the inventory status based on the pick-up and drop-off consistency result. The communication and collaboration interface unit is used to communicate bidirectionally with the operation and maintenance robot and / or the operation and maintenance management platform to send the pick-up and place task plan and pick-up and place guidance information to the operation and maintenance robot, and to receive execution feedback. The task orchestration unit is configured to pre-occupy the target item identifier and / or target storage location identifier after the authorization verification is passed, and release the pre-occupancy after the transaction and ledger unit completes the commit update.

2. The storage device according to claim 1, characterized in that, The storage resource management unit maintains the topological relationships and conflict constraints of storage locations, and the task orchestration unit determines the target storage location identifier and / or generates the pick-and-place action sequence based on the topological relationships and conflict constraints.

3. The storage device according to claim 1, characterized in that, The task orchestration unit performs two-stage orchestration for pick-up and drop-off requests: the first stage generates a set of candidate target storage locations and performs pre-occupancy; The second stage involves determining the final target storage location identifier and solidifying the retrieval and placement task plan after receiving the results of the pre-retrieval and placement verification.

4. The storage device according to claim 1, characterized in that, The permission and policy control unit is configured to generate a short-term authorization token corresponding to a single pick-up and drop-off request. The short-term authorization token includes at least the request number, robot identifier, target item identifier, target quantity, and validity period. The communication and collaboration interface unit is only allowed to issue pick-up and drop-off task plans and / or receive pick-up and drop-off completion confirmations when the short-term authorization token is valid.

5. The storage device according to claim 1, characterized in that, The state perception and verification unit includes a multi-source perception fusion subunit, which is used to fuse at least two different types of perception data to output the fusion judgment result of the target item's in-situ status, quantity change and / or abnormal status.

6. The storage device according to claim 5, characterized in that, The multi-source sensing fusion subunit calculates the confidence level of the fusion determination result and triggers at least one of the following processing strategies when the confidence level is lower than the threshold: re-acquisition, switching to an alternative verification path, or upgrading to manual review.

7. The storage device according to claim 1, characterized in that, The transaction and ledger unit is configured to establish an atomic transaction boundary for each take-up and release request, and record an inventory snapshot before take-up and release, an inventory snapshot after take-up and release, verification results and exception codes within the transaction boundary to support rollback updates and audit traceability.

8. The storage device according to claim 1, characterized in that, The transaction and ledger unit adopts an idempotent commit mechanism, which performs deduplication processing on duplicate pick-up and drop-off completion confirmations based on the request number, and performs inventory status commit update only when not committed.

9. A storage method based on the storage device according to any one of claims 1 to 8, characterized in that, include: S1. Receive a pick-up and place-down request sent by the maintenance robot, wherein the pick-up and place-down request includes at least the target item identifier and the target quantity; S2. Perform authorization verification on the pick-up and drop-off request, and perform pre-occupancy on the target item identifier and / or candidate target storage location identifier after the verification is passed; S3. Generate a retrieval and placement task plan based on the storage resource set, and send the retrieval and placement task plan and retrieval and placement guidance information to the operation and maintenance robot; S4. Collect sensing data before, during and / or after the pick-and-place process and output pick-and-place consistency results; S5. Write the pick-up and drop-off request into the inventory transaction record, and perform a commit update or rollback update on the inventory status based on the pick-up and drop-off consistency result. S6. Release the pre-occupied space and generate a ledger record, then send the result back to the operation and maintenance robot and / or the operation and maintenance management platform.

10. The method according to claim 9, characterized in that, When at least one of the following anomalies is detected: verification inconsistency, pre-occupancy timeout, robot execution failure, or communication interruption, an anomaly handling process is triggered. The anomaly handling process includes at least: revoking the short-term authorization token, freezing the target item identifier and / or the target storage location identifier, performing a rollback update, and generating an alarm event.

Images