Method and system for collaborative processing of unmanned aerial vehicle cruise line of converter station based on digital model

Abstract

The embodiment of the application provides a kind of based on digital model's converter station unmanned plane cruise route collaborative processing method and system, obtain network knowledge learning data sequence including multiple template collaborative control data and each prior abnormality mark corresponding to multiple template collaborative control data;Load basic exception label diagnostic network, and the basic exception label diagnostic network is the exception label diagnostic network generated by knowledge learning long short-term memory network model based on initialization weight parameter;Based on knowledge domain self-adaptability training, the prediction unit of the basic exception label diagnostic network is learned by network knowledge learning data sequence, and the exception analysis network is obtained;Prediction unit includes one or more multilayer perceptron;According to exception analysis network, abnormal node positioning is carried out, and the high-performance exception analysis network is generated using the technical solution, so that subsequent abnormal node positioning accurate, fast can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of AI, in particular to a digital model-based coordinated processing method and system for UAV cruise routes of a converter station. BACKGROUND

[0002] Artificial intelligence, cloud computing, big data, the Internet of Things, mobile Internet and other new-generation information technologies are gradually becoming powerful engines for intelligent life, and are greatly promoting the innovation and development of unmanned aerial vehicle technology. Autonomous patrol route planning technology, as one of the core technologies of unmanned aerial vehicle systems, can replace human planning and decision-making for flight tasks in complex dynamic environments, effectively avoid unsafe control behaviors of ground station operators, and share their task work pressure so that the crew can focus their efforts on key tasks. However, in reality, the increase in unmanned aerial vehicle operation tasks and the increasing complexity of flight environments, how to detect unmanned aerial vehicle cruise route coordination control data in real time, quickly and accurately find abnormal points, is a technical effect that needs to be solved in the field. SUMMARY

[0003] Therefore, the purpose of the embodiments of the present application is to provide a digital model-based coordinated processing method and system for UAV cruise routes of a converter station, obtain a network knowledge learning data sequence including a plurality of template coordination control data and a plurality of template coordination control data corresponding to each prior abnormal identifier; load a basic abnormal label diagnosis network, the basic abnormal label diagnosis network is an abnormal label diagnosis network generated by training a long short-term memory network model based on an initial weight parameter; based on knowledge domain adaptability training, train the prediction unit of the basic abnormal label diagnosis network through the network knowledge learning data sequence, to obtain an abnormal analysis network; the prediction unit includes one or more multilayer perceptrons; and locate the abnormal node according to the abnormal analysis network, the present technical solution generates a high-performance abnormal analysis network, so that subsequent abnormal node positioning is accurate and fast.

[0004] According to an aspect of the embodiments of the present application, a digital model-based coordinated processing method and system for UAV cruise routes of a converter station are provided, the method comprising:

[0005] obtaining a network knowledge learning data sequence, the network knowledge learning data sequence including a plurality of template coordination control data and a plurality of template coordination control data corresponding to each prior abnormal identifier;

[0006] loading a basic abnormal label diagnosis network, the basic abnormal label diagnosis network being an abnormal label diagnosis network generated by training a long short-term memory network model based on an initial weight parameter;

[0007] training the prediction unit of the basic abnormal label diagnosis network based on knowledge domain adaptivity training, and obtaining the abnormal analysis network through the network knowledge learning data sequence;

[0008] locating the abnormal node according to the abnormal analysis network.

[0009] In an alternative implementation, the training the prediction unit of the basic abnormal label diagnosis network based on knowledge domain adaptivity training, and obtaining the abnormal analysis network through the network knowledge learning data sequence, comprises:

[0010] locking the structure parameter information of the knowledge domain adaptivity unit of the basic abnormal label diagnosis network, and initializing the structure parameter information of the prediction unit to obtain a pending abnormal label diagnosis network; wherein the knowledge domain adaptivity unit is a structure layer combination composed of all structure units of the basic abnormal label diagnosis network except the prediction unit;

[0011] training the pending abnormal label diagnosis network according to the network knowledge learning data sequence to obtain the abnormal analysis network.

[0012] In an alternative implementation, the training the prediction unit of the basic abnormal label diagnosis network based on knowledge domain adaptivity training, and obtaining the abnormal analysis network through the network knowledge learning data sequence, comprises:

[0013] configuring a first training index parameter and a second training index parameter; wherein the first training index parameter represents the knowledge domain adaptivity parameter of the knowledge domain adaptivity unit of the basic abnormal label diagnosis network, and the second training index parameter represents the loss function influence weight of the prediction unit, and the first training index parameter is smaller than the second training index parameter;

[0014] training the pending abnormal label diagnosis network according to the network knowledge learning data sequence, the first training index parameter and the second training index parameter to obtain the abnormal analysis network.

[0015] In an alternative implementation, after training the prediction unit of the basic abnormal label diagnosis network based on knowledge domain adaptivity training, and obtaining the abnormal analysis network through the network knowledge learning data sequence, the method further comprises:

[0016] obtaining a network verification data sequence, wherein the network verification data sequence comprises a plurality of template cooperative control data and a prior abnormal identifier corresponding to each template cooperative control data in the plurality of template cooperative control data;

[0017] The abnormal analysis performance of the abnormal analysis network is verified according to the network verification data sequence.

[0018] In an alternative embodiment, before the abnormal analysis network is obtained by training the prediction unit of the basic abnormal label diagnosis network through the network knowledge learning data sequence based on knowledge domain self-adaptivity training, the method further comprises:

[0019] The multiple converter station unmanned aerial vehicle cruise line cooperative control data are subjected to principal component analysis to obtain a first cooperative control knowledge vector corresponding to the multiple converter station unmanned aerial vehicle cruise line cooperative control data;

[0020] The first cooperative control knowledge vector corresponding to the multiple converter station unmanned aerial vehicle cruise line cooperative control data is subjected to heuristic search to obtain a second cooperative control knowledge vector corresponding to the multiple converter station unmanned aerial vehicle cruise line cooperative control data;

[0021] The dimension of the second cooperative control knowledge vector corresponding to the multiple converter station unmanned aerial vehicle cruise line cooperative control data is converted into a target dimension, and the target dimension is a dimension of data adapted to the network loading part of the abnormal analysis network.

[0022] In an alternative embodiment, the abnormal node positioning according to the abnormal analysis network comprises:

[0023] The candidate unmanned aerial vehicle cruise line cooperative control data are obtained, and the candidate unmanned aerial vehicle cruise line cooperative control data include multiple task converter station unmanned aerial vehicle cruise line cooperative control data; and the candidate unmanned aerial vehicle cruise line cooperative control data are input into the abnormal analysis network;

[0024] According to the output of the abnormal analysis network, an identification of an abnormal event in the candidate unmanned aerial vehicle cruise line cooperative control data is obtained.

[0025] In an alternative embodiment, the basic abnormal label diagnosis network is a Monte Carlo network model.

[0026] According to another aspect of the embodiment of the application, a converter station unmanned aerial vehicle cruise line cooperative processing method and system based on a digital model are provided, and the system comprises:

[0027] The acquisition module is configured to acquire a network knowledge learning data sequence, and the network knowledge learning data sequence includes multiple template cooperative control data and a prior abnormal identification corresponding to each template cooperative control data in the multiple template cooperative control data;

[0028] The loading module is configured to load a basic exception label diagnosis network, which is an exception label diagnosis network generated by training a long short-term memory network model based on an initialized weight parameter;

[0029] The learning module is configured to train a prediction unit of the basic exception label diagnosis network by using the network knowledge learning data sequence based on knowledge domain adaptability training, to obtain an exception analysis network. The prediction unit includes one or more multilayer perceptrons.

[0030] The positioning module is configured to locate an abnormal node based on the exception analysis network.

[0031] According to another aspect of the embodiments of the present application, an electronic device is provided, which includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete the communication among each other through the communication bus; the memory is used to store a computer program; and the processor is used to execute the computer program, so as to implement the steps of the method for processing the UAV cruise route of the converter station based on the digital model.

[0032] According to another aspect of the embodiments of the present application, a readable storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the method for processing the UAV cruise route of the converter station based on the digital model can be executed.

[0033] In order to make the above object, features and advantages of the embodiments of the present application more apparent and understandable, the following will be described in detail with reference to the embodiments and in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some of the embodiments of the present application, and therefore should not be considered as a limitation to the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0035] Figure 1 Fig. 1 shows a component schematic diagram of a server provided by the embodiments of the present application;

[0036] Figure 2 Fig. 2 shows a flow schematic diagram of the method and system for processing the UAV cruise route of the converter station based on the digital model provided by the embodiments of the present application;

[0037] Figure 3 Fig. 3 shows a functional module block diagram of the system for processing the UAV cruise route of the converter station based on the digital model provided by the embodiments of the present application. Detailed Implementation

[0038] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0039] The terms “first,” “second,” “third,” etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0040] Figure 1 An exemplary component diagram of server 100 is shown. Server 100 may include one or more processors 104, such as one or more central processing units (CPUs), each of which may implement one or more hardware threads.

[0041] Server 100 may also include any storage medium 106 for storing information of any kind, such as code, settings, data, etc. Without limitation, storage medium 106 may include any type of RAM, any type of ROM, flash memory device, hard disk, optical disk, etc. More generally, any storage medium can use any technology to store information. Furthermore, any storage medium may provide volatile or non-volatile retention of information. Furthermore, any storage medium may represent a fixed or removable component of server 100. In one case, when processor 104 executes associated instructions stored in any storage medium or combination of storage media, server 100 may perform any operation of the associated instructions. Server 100 also includes one or more drive units 108 for interacting with any storage medium, such as hard disk drive units, optical disk drive units, etc.

[0042] The server 100 also includes input / output 110 (I / O) for receiving various inputs (via input units 112) and for providing various outputs (via output units 114). One particular output mechanism can include a presentation device 116 and associated graphical user interface (GUI) 118. The server 100 can also include one or more network interfaces 120 for exchanging data with other devices via one or more communication units 122. One or more communication buses 124 couple the above-described components together.

[0043] The communication units 122 can be implemented in any manner, such as through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication units 122 can include any combination of hardwired links, wireless links, router, gateway functionality, name servers 100, etc., governed by any protocol or combination of protocols.

[0044] Figure 2 A flowchart of a method and system for collaborative processing of a UAV cruise route of a converter station based on a digital model is shown. The method and system can be executed by a server 100 as shown in FIG. 1. The detailed steps of the method and system are described as follows. Figure 1

[0045] At step S110, a network knowledge learning data sequence is obtained. The network knowledge learning data sequence includes a plurality of template collaborative control data and a prior abnormality identifier corresponding to each template collaborative control data in the plurality of template collaborative control data.

[0046] At step S120, a basic abnormality label diagnosis network is loaded. The basic abnormality label diagnosis network is an abnormality label diagnosis network generated by training a long short-term memory network model based on an initialized weight parameter.

[0047] At step S130, based on knowledge domain adaptability training, a prediction unit of the basic abnormality label diagnosis network is trained by using the network knowledge learning data sequence to obtain an abnormality analysis network. The prediction unit includes one or more multilayer perceptrons.

[0048] At step S140, an abnormality node is located according to the abnormality analysis network.

[0049] ​Based on the above steps, the embodiment obtains a plurality of template cooperative control data and each prior abnormality identifier corresponding to the plurality of template cooperative control data in the network knowledge learning data sequence; loads a basic abnormality label diagnosis network, which is an abnormality label diagnosis network generated by training a long short-term memory network model based on an initial weight parameter; trains a prediction unit of the basic abnormality label diagnosis network through the network knowledge learning data sequence based on knowledge domain adaptability training to obtain an abnormality analysis network; the prediction unit includes one or more multilayer perceptrons; and performs abnormal node positioning according to the abnormality analysis network. The technical solution generates a high-performance abnormality analysis network, so that subsequent abnormal node positioning is accurate and fast.

[0050] In an alternative embodiment, after the first template system encryption control data of the system abnormality analysis network is obtained, the following steps are further included:

[0051] The first template system encryption control data of the system abnormality analysis network is subjected to regularized feature coding to generate the first template system encryption control data after regularized feature coding;

[0052] The first template system encryption control data of the system abnormality analysis network is loaded into a scrambling control feature to generate template scrambling load data, including:

[0053] The first template system encryption control data after regularized feature coding is loaded into a scrambling control feature to generate template scrambling load data.

[0054] In an alternative embodiment, the first template system encryption control data of the system abnormality analysis network is loaded into a scrambling control feature to generate template scrambling load data, including:

[0055] The first template system encryption control data of the system abnormality analysis network is loaded into a scrambling control feature based on heuristic search transmission to generate template scrambling load data.

[0056] In an alternative embodiment, the template scrambling load data is subjected to scrambling restoration to generate the second template system encryption control data of the system abnormality analysis network, including:

[0057] Obtaining scrambling feature data loaded into the template scrambling load data;

[0058] The template scrambling load data and the scrambling feature data loaded into the template scrambling load data are input into an AI neural network to generate the second template system encryption control data of the system abnormality analysis network.

[0059] In an alternative implementation, the system anomaly analysis network with initialized weight parameters is optimized based on the first template system encryption control data of the system anomaly analysis network and the second template system encryption control data of the system anomaly analysis network, to generate an optimized system anomaly analysis network, including:

[0060] The active learning is performed based on the first template system encryption control data of the system anomaly analysis network and the second template system encryption control data of the system anomaly analysis network, to generate the basic network weight information.

[0061] The system anomaly analysis network with initialized weight parameters is optimized based on the basic network weight information, to generate an optimized system anomaly analysis network.

[0062] In an alternative implementation, the active learning is performed based on the first template system encryption control data of the system anomaly analysis network and the second template system encryption control data of the system anomaly analysis network, to generate the basic network weight information, including:

[0063] The feature extraction is performed on the first template system encryption control data of the system anomaly analysis network and the second template system encryption control data of the system anomaly analysis network based on the active learning algorithm, to generate the basic network weight information.

[0064] In an alternative implementation, after the system anomaly analysis network with initialized weight parameters is optimized based on the first template system encryption control data of the system anomaly analysis network and the second template system encryption control data of the system anomaly analysis network, to generate an optimized system anomaly analysis network, further including:

[0065] Obtaining system encryption control data to be analyzed;

[0066] Loading the system encryption control data to be analyzed into the optimized system anomaly analysis network, to generate weight values of at least one abnormal encryption operation;

[0067] Determining abnormal encryption operation analysis data based on the weight values of the abnormal encryption operations.

[0068] Figure 3A functional module diagram of the system 200 for processing a UAV cruising route of a converter station based on a digital model according to an embodiment of the present application is shown. The functions implemented by the system 200 for processing a UAV cruising route of a converter station based on a digital model can correspond to the steps performed by the method described above. The system 200 for processing a UAV cruising route of a converter station based on a digital model can be understood as the server 100 described above, or a processor of the server 100, or a component independent of the server 100 or the processor and controlled by the server 100 to implement the functions of the present application. Figure 3 As shown, the functions of each functional module of the system 200 for processing a UAV cruising route of a converter station based on a digital model are described in detail below.

[0069] The acquisition unit is configured to acquire first template system encryption control data of the system anomaly analysis network, wherein the first template system encryption control data carries corresponding abnormal encryption operation tag data.

[0070] The first generation unit is configured to load the first template system encryption control data of the system anomaly analysis network to a scrambling control feature to generate template scrambling load data.

[0071] The second generation unit is configured to scramble and restore the template scrambling load data to generate second template system encryption control data of the system anomaly analysis network.

[0072] The optimization unit is configured to perform network optimization on the system anomaly analysis network with an initialized weight parameter according to the first template system encryption control data of the system anomaly analysis network and the second template system encryption control data of the system anomaly analysis network to generate an optimized system anomaly analysis network.

[0073] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the system, device and unit described above can refer to the corresponding processes in the foregoing method embodiments, which will not be described herein.

[0074] It is obvious for those skilled in the art that the present application is not limited to the details of the foregoing exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present application is defined by the appended claims rather than the foregoing description, and all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application.

Claims

1. A collaborative processing method for UAV cruise routes at converter stations based on digital models, characterized in that, The method includes: Obtain a network knowledge learning data sequence, the network knowledge learning data sequence including multiple template collaborative control data and prior anomaly identifiers corresponding to each template collaborative control data in the multiple template collaborative control data; Load the basic anomaly label diagnosis network, which is an anomaly label diagnosis network generated by training a long short-term memory network model based on initialized weight parameters; Based on knowledge domain adaptive training, the prediction unit of the basic anomaly label diagnosis network is trained using the network knowledge learning data sequence to obtain an anomaly analysis network; the prediction unit includes one or more multilayer perceptrons. Anomaly node location is performed based on the aforementioned anomaly analysis network; The knowledge domain-based adaptive training, which trains the prediction unit of the basic anomaly label diagnosis network using the network knowledge learning data sequence to obtain the anomaly analysis network, includes: The structural parameter information of the knowledge domain adaptive unit of the basic anomaly label diagnosis network is locked, and the structural parameter information of the prediction unit is initialized to obtain the anomaly label diagnosis network to be determined; wherein, the knowledge domain adaptive unit is a combination of structural layers composed of all structural units in the basic anomaly label diagnosis network except for the prediction unit. The anomaly analysis network is obtained by training the network knowledge learning data sequence on the network to determine the anomaly label diagnosis network.

2. The method for collaborative processing of UAV cruise routes at converter stations based on digital models according to claim 1, characterized in that, The knowledge domain-based adaptive training, which trains the prediction unit of the basic anomaly label diagnosis network using the network knowledge learning data sequence to obtain the anomaly analysis network, includes: Configure a first training metric parameter and a second training metric parameter; wherein, the first training metric parameter represents the knowledge domain adaptive parameter of the knowledge domain adaptive unit of the basic anomaly label diagnosis network, and the second training metric parameter represents the weight of the loss function of the prediction unit, and the first training metric parameter is smaller than the second training metric parameter; The anomaly analysis network is obtained by training the network knowledge learning data sequence, the first training index parameter, and the second training index parameter.

3. The method for collaborative processing of UAV cruise routes at converter stations based on digital models according to claim 1, characterized in that, After training the prediction unit of the basic anomaly label diagnosis network using the network knowledge learning data sequence based on knowledge domain adaptive training to obtain the anomaly analysis network, the method further includes: Obtain a network verification data sequence, the network verification data sequence including multiple template collaborative control data and a priori anomaly identifiers corresponding to each template collaborative control data in the multiple template collaborative control data; The anomaly analysis performance of the anomaly analysis network is verified based on the network verification data sequence.

4. The method for collaborative processing of UAV cruise routes at converter stations based on digital models according to claim 1, characterized in that, Before training the prediction unit of the basic anomaly label diagnosis network using the network knowledge learning data sequence based on knowledge domain adaptive training to obtain the anomaly analysis network, the method further includes: Principal component analysis was performed on the collaborative control data of the UAV cruise routes of the multiple converter stations to obtain the first collaborative control knowledge vector corresponding to the collaborative control data of the UAV cruise routes of each converter station. A heuristic search is performed on the first collaborative control knowledge vector corresponding to the collaborative control data of the UAV cruise route of each converter station to obtain the second collaborative control knowledge vector corresponding to the collaborative control data of the UAV cruise route of each converter station. The dimensions of the second collaborative control knowledge vector corresponding to the collaborative control data of the UAV cruise routes of each converter station are converted into target dimensions. The target dimensions are the dimensions of the data adapted by the network loading part of the anomaly analysis network.

5. The method for collaborative processing of UAV cruise routes at converter stations based on digital models according to claim 1, characterized in that, The step of locating abnormal nodes based on the anomaly analysis network includes: Acquire candidate UAV cruise route collaborative control data, which includes collaborative control data of converter station UAV cruise routes for multiple tasks; input the candidate UAV cruise route collaborative control data into the anomaly analysis network; Based on the output of the anomaly analysis network, the identifiers of anomalous events in the collaborative control data of the candidate UAV cruise routes are obtained.

6. The method for collaborative processing of UAV cruise routes at converter stations based on digital models according to any one of claims 1-5, characterized in that, The basic anomaly labeling diagnostic network is a Monte Carlo network model.

7. A collaborative processing system for UAV cruise routes at converter stations based on digital models, characterized in that, The system includes: The acquisition module is used to acquire a network knowledge learning data sequence, which includes multiple template collaborative control data and a priori anomaly identifiers corresponding to each template collaborative control data in the multiple template collaborative control data. The loading module is used to load the basic anomaly label diagnosis network, which is an anomaly label diagnosis network generated by training a long short-term memory network model based on the initialized weight parameters. The learning module is used for knowledge domain adaptive training. It trains the prediction unit of the basic anomaly label diagnosis network through the network knowledge learning data sequence to obtain the anomaly analysis network. The prediction unit includes one or more multilayer perceptrons. The positioning module is used to locate abnormal nodes based on the anomaly analysis network; The knowledge domain-based adaptive training, which trains the prediction unit of the basic anomaly label diagnosis network using the network knowledge learning data sequence to obtain the anomaly analysis network, includes: The structural parameter information of the knowledge domain adaptive unit of the basic anomaly label diagnosis network is locked, and the structural parameter information of the prediction unit is initialized to obtain the anomaly label diagnosis network to be determined; wherein, the knowledge domain adaptive unit is a combination of structural layers composed of all structural units in the basic anomaly label diagnosis network except for the prediction unit. The anomaly analysis network is obtained by training the network knowledge learning data sequence on the network to determine the anomaly label diagnosis network.

8. An electronic device, characterized in that, include: The processor, communication interface, memory, and communication bus are provided, wherein the processor, communication interface, and memory communicate with each other via the communication bus. The memory is used to store computer programs; the processor is used to execute the computer programs to implement the steps of the collaborative processing method for the UAV cruise route of the converter station based on the digital model as described in any one of claims 1-6.

9. A readable storage medium, characterized in that, The readable storage medium stores a computer program that, when executed by a processor, can perform the steps of the digital model-based UAV cruise route collaborative processing method for converter stations as described in any one of claims 1-6.

Images