A data processing method and an electronic device

Abstract

The application is suitable for the technical field of fault simulation, and provides a data processing method and an electronic device, which comprises: acquiring scene description information; the scene description information is used for describing a training scene, the training scene at least comprises a device fault, and the scene description information corresponding to different training scenes is different; generating a fault seed parameter according to the scene description information; the fault seed parameter is a parameter for triggering the training scene; generating a training scene according to the fault seed parameter; the training scene is used for training the operation behavior of a tested person on a device in the training scene. The data processing method and the electronic device provided by the application can provide a dynamic training scene for training.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of fault simulation, and particularly relates to a data processing method and an electronic device. BACKGROUND

[0002] At present, training scenes under equipment faults are mostly several typical scenes preset and limited, for example, several typical training scenes are stored in a training database, and the training scenes provided by the database are relatively single. After being trained based on the single training scene, a tested person is difficult to cope with a sudden training scene or often responds not timely enough to the sudden training scene. SUMMARY

[0003] Embodiments of the application provide a data processing method and an electronic device method, which can provide dynamic training scenes.

[0004] In a first aspect, embodiments of the application provide a data processing method, including: obtaining scene description information; the scene description information is used to describe a training scene, the training scene at least includes equipment failure, and corresponding scene description information is different under different training scenes; generating a fault seed parameter according to the scene description information; the fault seed parameter is a parameter for triggering the training scene to be generated; and generating the training scene according to the fault seed parameter; the training scene is used to train an operation behavior of a tested person on the equipment under the training scene.

[0005] Optionally, the generating the fault seed parameter according to the scene description information includes: inputting the scene description information into a generator to obtain the fault seed parameter; the generator is obtained by training a training sample, the training sample includes scene description information samples under different training scene samples, fault seed parameter samples corresponding to the scene description information samples, and association description information between the scene description information samples and the fault seed parameter samples, the association description information is used to describe an association relationship between the scene description information samples and the fault seed parameter samples; and the generating the training scene according to the fault seed parameter includes: inputting the fault seed parameter into a renderer to generate the training scene; the renderer is used to generate audio and / or video generated by equipment running under the training scene.

[0006] Optionally, the scene description information comprises target description sub-information of the device in multiple running stages, and the multiple running stages have a running sequence; the obtaining of the scene description information comprises: obtaining the multiple target description sub-information; the multiple target description sub-information are respectively used for describing multiple sub-training scenes in the training scene, and the multiple target description sub-information are used for respectively generating fault seed parameters in the sub-training scenes corresponding to the multiple running stages; and the generating of the training scene according to the fault seed parameters comprises: generating the sub-training scenes in the multiple running stages in sequence according to the fault seed parameters in the sub-training scenes corresponding to the multiple running stages in the running sequence of the multiple running stages.

[0007] Optionally, each of the running stages corresponds to multiple description sub-information; the obtaining of the multiple target description sub-information comprises: obtaining the multiple description sub-information corresponding to each of the running stages; and screening the target description sub-information from the multiple description sub-information as the description information of the sub-training scene in the running stage; wherein the sub-training scenes described by the description sub-information in the multiple running stages are combined to form a training scene which is different from the presented historical training scene.

[0008] Optionally, the obtaining of the target description sub-information comprises: obtaining the target description sub-information according to multiple condition vectors in the case that the condition vectors input into multiple condition input boxes are received; the condition vectors are used for representing basic information of the sub-training scene, the multiple condition input boxes are associated with each other, and the condition vector input into a previous condition input box is used for limiting the range of the condition vector input into a next condition input box.

[0009] Optionally, the obtaining of the scene description information comprises: screening, from scene description information of multiple training scenes, scene description information of a training scene with a training probability less than a first preset probability and / or an accuracy less than a second preset probability; the training probability is used for representing a probability of displaying the training scene, and the accuracy is used for representing an accuracy of an operation behavior of the measured personnel in the training scene.

[0010] Optionally, after the training scene is generated, the method further comprises: obtaining operation behavior data of the measured personnel in the training scene; determining a target operation score of the measured personnel according to the operation behavior data; and the target operation score is positively correlated with operation accuracy and / or timeliness of the measured personnel.

[0011] Optionally, the operation behavior data comprises the operation behavior and a time point at which the operation behavior is generated; and the determining the target operation score of the tested personnel according to the operation behavior data comprises: obtaining a first operation score according to an accuracy rate of the operation behavior, wherein the accuracy rate of the operation behavior is used to represent a proportion of correct operation behaviors in the operation behaviors performed by the tested personnel, and the operation behaviors are the number of preset operation behaviors expected in the training scenario; obtaining a second operation score according to a time deviation between the time point of the operation behavior and a preset time point, wherein the preset time point is used to represent a time point at which the correct operation behavior occurs in the training scenario; and obtaining the target operation score according to the first operation score and / or the second operation score.

[0012] Optionally, the operation behavior data further comprises a target decision made by the tested personnel after performing the operation behavior; and after the training scenario is generated, the method further comprises: generating an analysis report according to a contribution degree of the operation behavior performed by the tested personnel before making the target decision to the target decision, wherein the contribution degree is used to represent a contribution of the operation behavior to promoting the tested personnel to make the target decision, and the analysis report is used to represent a cause analysis of the tested personnel making the target decision.

[0013] In a second aspect, an electronic device is provided, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the data processing method provided in the present disclosure when executing the computer program.

[0014] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0015] The fault seed parameters can be generated based on the scene description information, and the training scenarios expected to be described by the scene description information can be generated based on the fault seed parameters. In this process, on the one hand, if a non-typical training scenario outside the training database is to be generated, the fault seed parameters in the non-typical training scenario can be generated based on the scene description information in the non-typical training scenario, and then the non-typical training scenario is generated based on the fault seed parameters to train the tested personnel, so that the tested personnel can be repeatedly trained in various non-typical training scenarios in addition to the typical training scenario, thereby improving the coping ability of the tested personnel in the face of sudden equipment failures in the training scenario. On the other hand, the fault seed parameters are automatically generated based on the scene description information, and the training scenario is automatically generated based on the fault seed parameters, without manual programming, thereby saving labor costs. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0017] Figure 1 is a step flow chart of a data processing method provided by an embodiment of the present application;

[0018] Figure 2 is a step flow chart of a data processing method provided by an embodiment of the present application;

[0019] Figure 3 is a schematic diagram of a generator and a renderer processing data provided by an embodiment of the present application;

[0020] Figure 4 is a step flow chart of a data processing method provided by an embodiment of the present application;

[0021] Figure 5 is a step flow chart of a data processing method provided by an embodiment of the present application;

[0022] Figure 6 is a schematic diagram of describer information in different running stages provided by an embodiment of the present application;

[0023] Figure 7 is a step flow chart of a data processing method provided by an embodiment of the present application;

[0024] Figure 8 is a step flow chart of a data processing method provided by an embodiment of the present application;

[0025] Figure 9 is a step flow chart of a data processing method provided by an embodiment of the present application;

[0026] Figure 10 is a schematic diagram of a training stage of a measured person provided by an embodiment of the present application;

[0027] Figure 11 is a structural block diagram of a data processing apparatus provided by an embodiment of the present application. DETAILED DESCRIPTION

[0028] In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular architectures, techniques, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known methods, devices, circuits, and

[0029] It is to be understood that the terminology "includes", "has", "holds", "contains" and / or "comprising", "including", "having" and their conjugates, as used herein, means "including but not limited to", and not to the exclusion of any other term or aspect.

[0030] It is also to be understood that the terminology "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items, as well as to the possibility that one or more of the listed items can be excluded.

[0031] As used in the description of the application and the appended claims, the term "if' can be interpreted to mean "when" or "upon" or "in response to determining" or "in response to detecting" depending on the context. Similarly, the phrase "if it is determined" or "if [a described condition or event] is detected" can be interpreted to mean "upon determining" or "in response to determining" or "upon [the described condition or event] being detected" or "in response to [the described condition or event] being detected", depending on the context.

[0032] In addition, the terms "first", "second", "third", etc. as used in the description of embodiments herein and throughout the claims (if any) are not used to connote any relative importance but are simply used for purposes of distinction.

[0033] Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the application. Thus, the appearances of the phrases "in one embodiment" or "in an embodiment" or "in a various embodiment" or "in other embodiments" in various places throughout this specification are not necessarily referring to the same embodiment, unless otherwise specified. The terms "comprising", "including", "containing", "having" and their conjugates, as used throughout this specification, mean "including but not limited to", unless otherwise expressly specified.

[0034] Figure 1 is an exemplary embodiment of the present disclosure, which is used to interpret the steps of the data processing method, please refer to Figure 1As shown, the data processing method comprises the following steps:

[0035] S101, acquire scene description information.

[0036] The scene description information is used to describe a training scene, and describes a device fault expected to be presented in the training scene. The training scene is used to train / test the operation behavior of a person under test when facing a device fault. The corresponding scene description information is different in different training scenes. The device can be a rocket, an aircraft, or the like. The training scene presents at least one type of data, such as a device fault, a fault time period, a fault severity, and a subsystem of the device involved in the training scene.

[0037] For example, taking a rocket as the device, the scene description information includes "60S to 120S after the rocket takes off, the thrust of the first-stage engine of the rocket decreases, and the severity ranges from 10% to 15%". In this process, the training scene described by the scene description information includes the device fault "thrust decrease of the first-stage engine of the rocket", the fault time period "60S to 120S after the rocket takes off", the fault severity "10% to 15%", and the device subsystems involved, such as "propulsion system, electrical system, and control system".

[0038] S102, generate a fault seed parameter according to the scene description information.

[0039] The fault seed parameter is a parameter that triggers the training scene, and can also be understood as a cause of the device fault in the training scene. The existence of the fault seed parameter or the change of the fault seed parameter means that the fault seed parameter will induce certain device faults on the device.

[0040] For example, the fault seed parameter is a disturbance parameter that disturbs the device parameter. Taking a rocket as the device, the fault seed parameter is a disturbance parameter that causes "slight decrease of the turbine pump efficiency". For example, the turbine pump efficiency is originally 70%, and the disturbed turbine pump efficiency is 65%. Therefore, the fault seed parameter is a disturbance parameter that causes slight decrease of the turbine pump efficiency by 5%. After the fault seed parameter is injected into the device, the turbine pump efficiency will decrease slightly, and the thrust of the first-stage engine of the rocket will decrease.

[0041] S103, generate the training scene according to the fault seed parameter.

[0042] Optionally, the fault seed parameter can be injected into a simulation device model, and then the training scene can be simulated.

[0043] It can be understood that the device model obtained by simulation is essentially a constructor for characterizing the device itself. By inputting the fault seed parameter into the constructor, a training scene induced by the fault seed parameter can be generated.

[0044] In the related art, the instructor randomly selects a training scene from a database to train the person being tested. Since the training scenes in the database are relatively typical and single, on the one hand, if the person being tested repeatedly trains the typical and single training scenes in the database in advance, the person being tested can only cope with the training scenes in the database and cannot cope with unexpected training scenes outside the database. On the other hand, the preset training scenes in the database need to be programmed manually, which brings additional labor costs. When new training scenes need to be created or programmed in the database, a long period of time is also needed. The update speed of the new training scenes is slow, and it is difficult to keep up with the iterative updates of the device and the changes in the training tasks. Often, after the training task changes or the device is updated, the training scenes after the task changes or the device is updated have not been programmed.

[0045] Through the above technical solution, the fault seed parameter can be generated based on the scene description information, and the training scene expected to occur can be generated based on the fault seed parameter. In this process, on the one hand, if a non-typical training scene other than a typical training scene is to be generated, the fault seed parameter under the non-typical training scene can be generated based on the scene description information under the non-typical training scene, and then the non-typical training scene can be generated based on the fault seed parameter to train the person being tested, so that the person being tested can repeatedly train in various typical or non-typical training scenes, and the coping ability of the person being tested in the face of unexpected device faults can be improved. On the other hand, the fault seed parameter is automatically generated based on the scene description information, and the training scene is automatically generated based on the fault seed parameter, without manual programming, thereby saving labor costs.

[0046] Figure 2 With Figure 3 The example embodiments related to the present disclosure are used to interpret the fault seed parameter generated by the generator under the scene description information, and the training scene generated by the renderer based on the fault seed parameter, including the following steps:

[0047] S102-1, input the scene description information into the generator to obtain the fault seed parameter.

[0048] The generator can be a C-GAN (Conditional Generative Adversarial Network), the scene description information includes a plurality of dimensional condition vectors, the plurality of dimensional condition vectors have a correlation relationship in line with a physical law, the plurality of condition vectors C can be input to the trained generator G to predict the fault seed parameter; the fault seed parameter is input into the discriminator D to obtain the error between the fault seed parameter predicted by the generator G and the actual fault seed parameter, and the score of the predicted fault seed parameter is obtained according to the error, and the greater the error, the lower the score of the predicted fault seed parameter.

[0049] Optionally, the generator can be trained by using a plurality of different training samples, each set of training samples generates a different training scene, and the training samples include scene description information samples under the training scene samples, fault seed parameter samples corresponding to the scene description information samples, and association description information between the scene description information samples and the fault seed parameter samples.

[0050] The scene information description sample is used to describe the training scene sample, the fault seed parameter sample is a parameter sample used to trigger the generation of the training scene sample, the association description information between the scene description information sample and the fault seed parameter sample is used to describe the association relationship between the scene description information sample and the fault seed parameter sample, and the association relationship also indicates a potential physical law relationship between the fault seed parameter sample and the scene description information sample, for example, the association relationship indicates that the fault seed parameter sample will induce the device failure described by the scene description information sample to occur.

[0051] For example, if the scene description information sample in the training sample is "60S to 120S after the rocket takes off, there is an abnormal flame at the engine nozzle", and the fault seed parameter sample is "the disturbance parameter of valve leakage", then the association description information can be "the disturbance parameter of valve leakage will induce the valve leakage to increase, and further cause the abnormal flame at the engine nozzle", in this way, after the training sample is input into the generator, the generator can quickly learn the causal relationship or physical law relationship between the scene description information sample and the fault seed parameter sample, and then the generator can generate accurate fault seed parameters and fault seed parameters in line with the physical law relationship after receiving the scene description information.

[0052] S103-1, input the fault seed parameter into the renderer to generate the training scene.

[0053] The renderer is configured to generate audio and / or video produced by the device in the training scene. The audio includes sound effects produced by the device during operation, and prompt sounds of virtual characters, for example, flight sound effects of a rocket during operation, and prompt sounds of a test control officer or a weather officer, which interact with the testee through voice interaction, thereby enhancing the immersion and stress of the testee. The video includes flight special effects of the device during operation, and failure special effects of the device failure, which enhances the realism of the training test by presenting more realistic failure special effects.

[0054] The renderer includes a simulator and a rendering engine. The simulator is configured to create a three-dimensional device model in the rendering engine. The rendering engine can be a UE5 (Unreal Engine 5) rendering engine. The simulator is configured to create a three-dimensional device model in the rendering engine, and the rendering engine is configured to render the display effect of the device model in each operation stage, such as dynamic light and shadow, explosion, fragmentation, vibration, and the like, thereby creating the realism of the device failure.

[0055] For example, if the normal leakage area of the device is A, and the fault seed parameter is a disturbance parameter that increases the leakage area by 15%, the fault seed parameter can be injected into the three-dimensional device model, thereby perturbing the leakage area in the device model from A to A+A*15%, and modifying the leakage area from the normal area to the abnormal area. After the leakage area of the device is abnormal, the corresponding physical equation of the simulation model is coupled and nonlinear, so the deviation of the leakage area will be amplified, transmitted and coupled over time, and then present a chain reaction such as tail gas temperature rise, unstable combustion, and thrust decline, thereby evolving the real evolution process of the device failure. Finally, the real evolution process of the device failure is input into the rendering engine, thereby displaying the real evolution effect of the device failure, for example, if the rendering engine receives a tail gas temperature rise, the tail gas produced by the device model will change from orange to bright yellow; if it receives an unstable combustion, it will display the rate, size and noise intensity of the flame burning curve on the console of the device model; if it receives a thrust decline, it will make the flight acceleration and sound wave pitch of the device model change synchronously.

[0056] When training the generator, the scene description information sample and the fault seed parameter sample are originally considered to train the generator, so that the generator can learn the physical law relationship between the scene description information sample and the fault seed parameter sample. However, this training method is slow and has low accuracy.

[0057] By the technical solution, in addition to training the generator by the scene description information sample and the fault seed parameter sample, the generator is input with the potential association description information between the scene description information sample and the fault seed parameter sample, so that the generator quickly learns the potential association relationship between the scene description information sample and the fault seed parameter sample; secondly, the generator does not randomly manufacture the fault seed parameter under the scene description information, that is, the generator does not randomly manufacture the confusion on the device, the association relationship reflects the physical law, which reflects the fault seed parameter sample that causes the training scene sample to generate in the training scene sample described by the scene description information sample and conforms to the physical law, so that after the generator is trained based on the association description information, the generator can generate the fault seed parameter that conforms to the physical law and is accurate based on the scene description information; thirdly, the audio and / or video when the device fails in the training scene can be displayed in real time by the renderer, so that the measured person interacts with the audio displayed in real time in the training scene, and watches the rendering picture of the training scene through the video, increasing the immersion and reality of the measured person when training.

[0058] Figure 4 is an exemplary embodiment related to the present disclosure, which is an exemplary scheme for training the training scene of the paraphrase generation device in multiple running stages, including the following steps:

[0059] S101-1, obtaining multiple target description sub-information.

[0060] The scene description information includes description sub-information of the device in multiple running stages, and the multiple target description sub-information is used to describe multiple sub-training scenes in multiple training scenes. The multiple running stages have a running order, and the multiple sub-training scenes have a running order.

[0061] For example, the multiple running stages of the device in the training scene include three running stages of the first running stage, the second running stage and the third running stage.

[0062] The target description sub-information in the first running stage includes: 60 seconds to 120 seconds after the rocket takes off (i.e. the maximum dynamic pressure section); target fault: first-stage engine thrust drop; severity: 10%-15%.

[0063] The target description sub-information in the second running stage includes: 120 seconds to 170 seconds after the rocket takes off; target fault: first-stage engine mixture ratio abnormal drift; severity: mixture ratio offset +8%, thrust additional drop 3%.

[0064] The target descriptor information in the third operation stage includes: 170-250 seconds after the rocket takes off; target failure: local overheating of the second engine combustion chamber cooling channel; severity: wall temperature overrun by 12%, triggering passive thrust reduction by 8%.

[0065] It can be seen that the three operation stages are three stages with a running order in the device running process.

[0066] The plurality of target descriptor information is used to generate fault seed parameters in the sub-training scene corresponding to the plurality of operation stages respectively. For each target descriptor information, the target descriptor information can be input into the generator to generate the fault seed parameters in the operation stage corresponding to the target descriptor information. The fault seed parameters in different operation stages are different.

[0067] For example, the fault seed parameter in the first operation stage is a disturbance parameter causing the turbine pump efficiency to decrease, the fault seed parameter in the second operation stage is a disturbance parameter causing the gas generator fuel throttle valve opening to decrease, and the fault seed parameter in the third operation stage is a disturbance parameter causing the cooling flow coefficient to decrease.

[0068] S103-2, according to the running order of the plurality of operation stages, generates sub-training scenes in the plurality of operation stages according to the fault seed parameters in the sub-training scenes corresponding to the plurality of operation stages in sequence.

[0069] Optionally, after obtaining the fault seed parameters in the plurality of operation stages, the fault seed parameters can be input into the renderer in sequence, so as to sequentially render and display the sub-training scenes in the plurality of operation stages. The sub-training scene is a sub-training scene in the entire training scene.

[0070] For example, continuing the example in step S101-1, the fault seed parameter "disturbance parameter causing turbine pump efficiency to decrease" in the first operating phase can be input into the renderer to render a sub-training scene in the first operating phase, so that the testee can see on the display screen the failure phenomenon of the first-stage engine thrust of the rocket decreasing; when the first operating phase ends and the second operating phase arrives, the fault seed parameter "disturbance parameter causing the fuel throttle valve opening of the gas generator to decrease" in the second operating phase can be input into the renderer to render a sub-training scene in the second operating phase, so that the testee can see on the display screen the failure phenomenon of the first-stage engine mixture ratio abnormally drifting; when the second operating phase ends and the third operating phase arrives, the fault seed parameter "disturbance parameter causing the cooling flow coefficient to decrease" in the third operating phase can be input into the renderer to render a sub-training scene in the third operating phase, so that the testee can see on the display screen the failure phenomenon of the local overheating of the cooling channel of the second-stage engine combustion chamber of the rocket.

[0071] It can be understood that the target description sub-information of the device in the multiple operating phases can be randomly selected and combined by the system from the database, or can be the target description sub-information of the device in the multiple operating phases set by the test personnel (for example, the instructor) according to their own experience and in accordance with physical laws.

[0072] In the related art, the training sub-scenes of the device in the multiple operating phases obtained from the database are single and linear, and the training sub-scenes of the device in the multiple operating phases do not dynamically change. For example, if the test personnel selects training scene A, the device evolves from sub-training scene A to sub-training scene B and then to sub-training scene C, but does not evolve from sub-training scene A to sub-training scene E and then to sub-training scene F. The combination of these training sub-scenes is fixed and cannot dynamically change, which leads to the training of the testee being relatively fixed and single, and the testee cannot be trained to respond to unexpected training scenes.

[0073] Through the above technical solution, multiple target description sub-information for describing different sub-training scenes can be obtained, and multiple target description sub-information is used to generate sub-training scenes in multiple operating phases, and finally the sub-training scenes in multiple operating phases are combined into a complete training scene. In this process, any set target description sub-information can be input to change a certain sub-training scene in the training scene, or different sub-training scenes described by different target description sub-information can be combined to form a new training scene. The training scene obtained by the combination can dynamically change with the change of the target description sub-information, and is no longer a training scene that changes along a single trajectory. The change mode is various, which can better train the testee to respond to unexpected situations.

[0074] Figure 5 is an exemplary scheme involved in the above step S101-1, the exemplary embodiment is used to interpret the acquisition of training scenarios that have never been combined to train the testee, including the following sub-steps:

[0075] A1, acquiring a plurality of corresponding description sub-information under each running phase.

[0076] Wherein, different types of equipment failures can be generated under each running phase, therefore different description sub-information can be used to describe different types of equipment failures generated under each running phase.

[0077] For example, referring to Figure 6 , the corresponding description sub-information under the first running phase includes description sub-information A, description sub-information B and description sub-information C, these three description sub-information are respectively used to describe equipment failure A, equipment failure B and equipment failure C; the corresponding description sub-information of the second running phase includes description sub-information D, description sub-information E and description sub-information F, these three description sub-information are respectively used to describe equipment failure D, equipment failure E and equipment failure F; the corresponding description sub-information of the third running phase includes description sub-information G, description sub-information H and description sub-information I, these three description sub-information are respectively used to describe equipment failure G, equipment failure H and equipment failure I.

[0078] A2, from a plurality of the description sub-information, screening out the target description sub-information as the description information of the sub-training scene under the running phase.

[0079] Wherein, the sub-training scenes described by the description sub-information under a plurality of running phases are different after being combined to form a training scene, and the presented historical training scene is a training scene that has been displayed on the display screen and trained the testee before. It can also be understood that the target description sub-information combined under a plurality of running phases is a target description sub-information that has never been combined before.

[0080] For example, continuing the example in sub-step A1, the description sub-information A, description sub-information B and description sub-information C of the first running phase, the description sub-information D, description sub-information E and description sub-information F of the second running phase, and the description sub-information G, description sub-information H and description sub-information I of the third running phase are combined, which can produce 27 combinations, that is, 27 training scenes can be generated, but among the 27 training scenes, 26 training scenes have appeared as historical training scenes before, and the training scene described by the combination of description sub-information A, description sub-information D and description sub-information G has not appeared in the historical training scene, therefore the training scene described by the combination of description sub-information A, description sub-information D and description sub-information G can be used as the training scene of the testee this time.

[0081] Through the technical solution, the corresponding multiple description sub-information under multiple running stages can be obtained, and the target description sub-information is filtered from the multiple description sub-information, so that the training scene formed by combining the target description sub-information under multiple running stages is a historical training scene that has not appeared before. Therefore, the training scene filtered each time is a new training scene after combination, so that the training scene for training the next measured personnel is different from the training scene for training the last measured personnel, avoiding that the measured personnel knows the training scene in advance and performs targeted training, and further training the measured personnel to face the response ability to the sudden training scene. In addition, the measured personnel is trained by using the training scene with faults, which can lead to the thinking solidification of the measured personnel, and the response ability of the measured personnel to face the new training scene can also be improved after the measured personnel is trained by using the dynamically updated training scene.

[0082] Figure 7 The above step S101-1 involves an exemplary scheme, and the exemplary embodiment is used to interpret the exemplary scheme of obtaining the target description sub-information, including the following sub-steps:

[0083] B1, in the case of receiving the condition vector input into the multiple condition input boxes, obtaining the target description sub-information according to the multiple condition vectors.

[0084] The condition vectors input in the multiple condition input boxes are combined to represent the target description sub-information under a single running stage, and the target description sub-information under multiple running stages is combined to represent the scene description information under the entire training scene. The condition vector is used to represent the basic information of the sub-training scene, and the multiple condition input boxes are associated with each other. The condition vector input in the last condition input box is used to limit the range of the condition vector input in the next condition input box.

[0085] For example, taking the number of multiple condition input boxes as three, the first condition input box is "task stage", the second condition input box is "target fault", and the third condition input box is "severity". After the test personnel inputs "60 seconds to 120 seconds after the rocket takes off" in the first condition input box, multiple alternative target faults such as "first-stage engine thrust reduction", "fairing indentation", and "engine throttling" are displayed in the second condition input box. After the user selects the target fault "first-stage engine thrust reduction", the severity range of "1%~50%" is displayed in the third condition input box so that the user can select the severity.

[0086] In this process, after the condition vector input in the last condition input box is determined, the target fault displayed in the next condition input box is not all the device faults that can occur in the entire device, but the device fault output under the restriction of the condition vector input in the last condition input box, which makes the multiple condition vectors input in the multiple condition input boxes comply with the physical law. For example, if the first running stage "60 seconds to 120 seconds after the rocket takes off" is input in the first condition input box, the target fault under the second running stage will not be popped up in the second condition input box.

[0087] Through the above technical solution, after the test personnel input the condition vectors in the multiple condition input boxes, if the condition vector is input in the last condition input box, the alternative condition vector can be selected from the next condition input box. In this way, the combined multiple condition vectors obtained through multiple selections are not randomly combined, but comply with the physical law.

[0088] Figure 8 The above step S101 relates to an exemplary embodiment for releasing the training scene described by the scene description information of the training scene with small training probability or low accuracy rate as an exemplary scheme of the training scene, including the following steps:

[0089] S101-2, from the scene description information of the multiple training scenes, screen out the scene description information of the training scene with a training probability less than a first preset probability and / or an accuracy rate less than a second preset probability.

[0090] The training probability is used to represent the probability of displaying the training scene, which represents the proportion of the number of times the training scene is displayed on the display screen to the total number of displays. The smaller the training probability, the fewer the number of times the training scene appears. The training probability less than the first preset probability is used to represent that the training probability of the training scene is small.

[0091] The accuracy rate is used to represent the accuracy rate of the operation behavior of the measured personnel in the training scene, which represents the proportion of the correct operation behavior generated by the measured personnel to the total number of operation behaviors in the training scene. For example, in this training scene, the measured personnel needs to make 10 operation behaviors, but the measured personnel makes 9 operation behaviors, of which 8 operation behaviors are accurate and the remaining 1 is incorrect, so the accuracy rate is 80%. The accuracy rate less than the second preset probability is used to represent that the accuracy rate of the operation behavior of the measured personnel in the training scene is low.

[0092] By the technical solution, the training scene described by the scene description information with the training probability less than the first preset probability can be screened from the scene description information of the multiple training scenes to train the measured personnel, so that the measured personnel can be trained in the training scene with the small probability, and the coping ability of the measured personnel to the small probability device fault can be improved; the training scene described by the scene description information with the accuracy less than the second preset probability can be screened from the scene description information of the multiple training scenes to train the measured personnel, so that the measured personnel can be trained in the training scene with the frequent operation error and low accuracy, the measured personnel can be repeatedly trained in the training scene with the low accuracy, and the professional accomplishment of the measured personnel can be improved.

[0093] Figure 9 The example embodiment related to the present disclosure is used for generating the target operation score of the measured personnel after the training scene is generated, and an example scheme includes the following steps.

[0094] S104, operation behavior data of the measured personnel in the training scene is acquired.

[0095] The operation behavior data is the operation behavior data of the measured personnel in the case of facing the training scene displayed on the display screen of the electronic device. The operation behavior data includes each operation behavior of the measured personnel and the time point of each operation behavior. The operation behavior includes the instruction output by the measured personnel, which is not limited to the key instruction, mouse instruction, voice instruction and the like input on the terminal. It can be understood that the fault simulation effect of the whole device is displayed on the display screen of the terminal, and the terminal can be a computer, a tablet, a notebook and the like.

[0096] S105, the target operation score of the measured personnel is determined according to the operation behavior data.

[0097] Optionally, the first operation score can be obtained according to the accuracy of the operation behavior, the second operation score can be obtained according to the time deviation between the time point of the operation behavior and the preset time point, and the target operation score can be obtained according to the first operation score and / or the second operation score.

[0098] The accuracy of the operation behavior is used to represent the proportion of the correct operation behavior in the operation behavior quantity in the operation behavior of the measured personnel, and the operation behavior quantity is the quantity of the preset operation behavior expected in the training scene. The lower the accuracy of the operation behavior is, the lower the first operation score is.

[0099] Whether the operation behavior performed by the measured person belongs to the correct operation behavior can be determined by whether the operation behavior performed by the measured person conforms to the standard, if it conforms to the standard, it belongs to the correct operation behavior. For example, if the measured person should use the mouse or keyboard to input the instruction of "preparing to execute the emergency plan", and the correct operation behavior is to issue the voice instruction of "preparing to execute the emergency plan", so the measured person does not perform according to the correct operation standard, which belongs to the incorrect operation behavior.

[0100] The preset time point is used to represent the time point at which the correct operation behavior occurs in the training scene. The greater the time deviation between the time point of the operation behavior generated by the measured person and the preset time point of the correct operation behavior, the longer the response time of the measured person to generate the operation behavior, and the smaller the second operation score.

[0101] Optionally, according to the first operation score and the second operation score, a target operation score is obtained, including: taking the sum of the first operation score and the second operation score as the target operation score.

[0102] Optionally, the operation behavior data further includes a target decision made by the measured person after performing the operation behavior, and an analysis report can be generated according to the contribution degree of the operation behavior performed by the measured person before making the target decision to the target decision.

[0103] The contribution degree of the operation behavior made between the two adjacent decisions to the target decision in the two decisions is used to generate the analysis report, and the target decision in the two decisions is the next decision in the two decisions.

[0104] The analysis report includes the target operation score of the measured person, the cause analysis of making the target decision, the improvement suggestion, the time point of each decision, the time point of each operation behavior, etc.

[0105] The contribution degree of the operation behavior to the target decision includes positive contribution degree and negative contribution degree. The positive contribution degree represents that the generation of the operation behavior has a positive driving effect on the target decision, and the negative contribution represents that no operation behavior is generated or the generated operation behavior has a negative driving effect on the target decision.

[0106] For example, if the two adjacent decisions are "prepare to execute the emergency plan" and "continue to fly", and the operation behaviors generated by the measured personnel between the two adjacent decisions are obtained, for example, the operation behaviors include: a mouse instruction is generated for the continuous decline of the thrust curve, a voice instruction is generated for the parameter A alarm, and no instruction is issued for the fluctuation of parameter B within the normal range, it can be determined that the two operation behaviors contribute to the target decision "continue to fly" to generate an analysis report. The analysis report includes the cause analysis "it is found that the continue to fly decision made by the measured personnel is mainly driven by the two factors of continuous decline of the thrust curve and parameter A alarm, but is negatively interfered by the information that the fluctuation of parameter B is within the normal range, resulting in that the continue to fly decision made by the measured personnel is three seconds later than the correct continue to fly decision", and the improvement suggestion "cognition training for operation behavior of parameter B under abnormal working conditions should be strengthened".

[0107] In the related art, the test personnel writes an analysis report based on the performance of the operation behavior of the measured personnel, and the prepared analysis report depends on the level and state of each test personnel, is subjective, lacks objective data support, and it is difficult to evaluate an accurate analysis report. Moreover, the efficiency of manually writing an analysis report is low, the depth is insufficient, manual review is time-consuming and laborious, and usually only a few obvious decision nodes can be focused on, it is difficult to deeply mine the massive data (all operations, communications, system states) generated in the drill, and it is difficult to locate the root cause of the hidden decision-making error. In addition, the analysis report fed back by the test personnel is a qualitative analysis of the operation behavior, and the evaluation is good or bad, and the target operation score of the operation behavior cannot be analyzed.

[0108] Through the above technical solutions, in a first aspect, the target operation score of the measured personnel in this training scenario can be automatically obtained based on the accuracy and time point of the operation behavior of the measured personnel, so as to quantitatively analyze the good or bad of the operation behavior of the measured personnel; in a second aspect, an analysis report can be generated based on the contribution of the operation behavior made by the measured personnel to the target decision, which shows the cause analysis of the target decision made by the measured personnel, so as to deeply mine the root cause of the decision-making error; in a third aspect, the analysis report is automatically generated based on the contribution of the operation behavior of the measured personnel to the target decision, without manual compilation and review, and the objectivity is strong. The analysis report can accurately reflect the training level of the measured personnel.

[0109] It should be understood that the size of the serial number of each step in the above embodiments does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0110] In some scenarios, please refer to Figure 10As shown, taking the device as a rocket as an example, in the first stage (virtual-real mapping stage), the space launch site in the real physical world can be laser scanned and data modeled, and the scanned data is input into the UE5 simulator to obtain a three-dimensional device model and a launch site model where the device model is located. The UE5 simulator can be a UE5 high-fidelity digital twin simulation module. In the second stage (dynamic generation stage), the condition vector input by the test personnel (for example, the instructor) is input into the generator, the generator injects the fault seed parameter into the simulated device model, drives the device model, and renders the display effect of the driven device model through the UE5 engine. In the second stage, the fault seed parameter and the audio data and other interaction data generated by the AI intelligent partner are injected into the simulated device model. In the third stage (training and data collection stage), the measured personnel operate to generate multi-modal operation behavior data such as voice, keyboard input, mouse input, and physiological data, and store the multi-modal operation behavior data after collection. In the fourth stage (intelligent review stage), the operation behavior data is obtained, and the operation behavior data is input into the intelligent review evaluation module to generate an analysis report and propose improvement suggestions in the analysis report. The analysis report is displayed to the test personnel and the measured personnel, and the intelligent review evaluation module also feeds back the optimized parameters, such as the fault seed parameter, to the generator, so that the generator corrects based on the real fault seed parameter.

[0111] corresponding to the data processing method described in the above embodiment, Figure 11 A structure block diagram of the data processing apparatus 1100 provided by the embodiments of the present application is shown, and only parts related to the embodiments of the present application are shown for ease of illustration.

[0112] With reference to Figure 11 The data processing apparatus 1100 comprises:

[0113] The acquisition module 1110 is configured to acquire scene description information; the scene description information is used to describe a training scene, and the training scene at least includes a device fault; the scene description information corresponding to different training scenes is different;

[0114] The seed generation module 1120 is configured to generate a fault seed parameter according to the scene description information; the fault seed parameter is a parameter that triggers the training scene to occur;

[0115] The scene generation module 1130 is configured to generate the training scene according to the fault seed parameter; the training scene is used to train the operation behavior of the measured personnel on the device in the training scene.

[0116] Optionally, the seed generation module 1120 is further configured to input the scene description information into a generator to obtain the fault seed parameter, the generator being trained by training samples, the training samples including scene description information samples under different training scene samples, fault seed parameter samples corresponding to the scene description information samples, and association description information between the scene description information samples and the fault seed parameter samples, the association description information being used to describe an association relationship between the scene description information samples and the fault seed parameter samples; and the scene generation module 1130 is further configured to input the fault seed parameter into a renderer to generate the training scene, the renderer being used to generate audio and / or video generated by device running in the training scene.

[0117] Optionally, the scene description information includes target description sub-information of the device in multiple running stages, the multiple running stages having a running order; the obtaining module 1110 is further configured to obtain the multiple target description sub-information, the multiple target description sub-information being respectively used to describe multiple sub-training scenes in the training scene, and the multiple target description sub-information being used to respectively generate fault seed parameters in the sub-training scenes corresponding to the multiple running stages; and the scene generation module 1130 is further configured to generate the sub-training scenes in the multiple running stages in sequence according to the fault seed parameters in the sub-training scenes corresponding to the multiple running stages in the running order of the multiple running stages.

[0118] Optionally, each of the running stages corresponds to multiple description sub-information; the obtaining module 1110 is further configured to obtain the multiple description sub-information corresponding to each of the running stages; and the target description sub-information is selected from the multiple description sub-information as description information of the sub-training scene in the running stage; and the sub-training scenes described by the description sub-information in the multiple running stages are combined to form a training scene different from the presented historical training scene.

[0119] Optionally, the obtaining module 1110 is further configured to obtain the target description sub-information according to multiple condition vectors in the case that the condition vectors input into multiple condition input boxes are received, the condition vectors being used to represent basic information of the sub-training scene, the multiple condition input boxes being associated with each other, and a condition vector input into a condition input box being used to limit a range of a condition vector input into a next condition input box.

[0120] Optionally, the acquisition module 1110 is further configured to filter out the scene description information of training scenarios whose training probability is less than a first preset probability and / or whose accuracy is less than a second preset probability from the scene description information of multiple training scenarios; the training probability is used to characterize the probability of displaying the training scenario, and the accuracy is used to characterize the accuracy of the test subject's operational behavior in the training scenario.

[0121] Optionally, the data processing apparatus 1100 includes:

[0122] The behavior acquisition module is used to acquire the operational behavior data of the test subject in the training scenario;

[0123] The scoring calculation module is used to determine the target operation score of the test subject based on the operation behavior data; the target operation score is positively correlated with the accuracy and / or timeliness of the test subject's operation.

[0124] Optionally, the operational behavior data includes the operational behavior and the time point at which the operational behavior occurred; the scoring calculation module is further configured to obtain a first operational score based on the accuracy rate of the operational behavior; the accuracy rate of the operational behavior is used to characterize the proportion of correct operational behaviors among the operational behaviors performed by the test subject, and the number of operational behaviors is the expected preset number of operational behaviors in the training scenario; a second operational score is obtained based on the time deviation between the time point of the operational behavior and the preset time point; the preset time point is used to characterize the time point at which the correct operational behavior occurs in the training scenario; and the target operational score is obtained based on the first operational score and / or the second operational score.

[0125] Optionally, the operational behavior data further includes a target decision, which is the decision made by the person being tested after performing the operational behavior; the data processing device 1100 includes:

[0126] The report generation module is used to generate an analysis report based on the contribution of the operational behaviors performed by the test subjects before making the target decision to the target decision; the contribution is used to characterize the contribution of the operational behaviors to prompting the test subjects to make the target decision, and the analysis report is used to characterize the reason analysis of the test subjects making the target decision.

[0127] It should be noted that the information interaction and execution process between the above-mentioned devices / units are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, and they will not be repeated here.

[0128] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the apparatus is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit or module in the embodiment can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit, and the integrated unit can be realized in the form of hardware or in the form of a software functional unit. In addition, the specific names of each functional unit or module are only for convenient distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0129] The embodiments of the present application further provide an electronic device, which comprises at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps in any of the method embodiments described above when executing the computer program.

[0130] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program, wherein the computer program is executable by a processor to implement the steps in any of the method embodiments described above.

[0131] The embodiments of the present application provide a computer program product, which, when running on an electronic device, enables the electronic device to implement the steps in any of the method embodiments described above.

[0132] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the present application can implement all or part of the processes in the above-mentioned embodiment methods through a computer program to instruct related hardware to complete, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium at least includes any entity or device capable of carrying the computer program code to an electronic device, a recording medium, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunications signal and a software distribution medium. For example, a U disk, a mobile hard disk, a magnetic disk or an optical disk, etc. In some jurisdictions, according to legislation and patent practice, the computer readable medium can not be an electrical carrier signal and a telecommunications signal.

[0133] The program code contained in the computer readable medium can be transmitted by any appropriate medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0134] The computer program code for executing the operation of the embodiments of the present application can be written in one or more programming languages or combinations thereof, including object oriented programming languages such as python, Java, Smalltalk, C++, and conventional procedural programming languages such as "C" language or similar programming languages. The program code can be executed completely on a user computer, partially on a user computer, as an independent software package, partially on a user computer and partially on a remote computer, or completely on a remote computer or server. In the case of remote computer, the remote computer can be connected to the user computer through any kind of network, including local area network (LAN) or wide area network (WAN), or can be connected to external computer (for example, through Internet service provider to connect through Internet).

[0135] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0136] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0137] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / equipment and method can be implemented in other ways. For example, the apparatus / equipment embodiments described above are merely schematic. The division of the modules or units is merely a logical function division. There can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual coupling or direct coupling or communication connection can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or in other forms.

[0138] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, i.e. they can be located in one place, or distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiments.

[0139] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A data processing method, characterized by, The method comprises: obtaining scene description information; the scene description information is used for describing a training scene, and the training scene at least comprises device failure, and the scene description information corresponding to different training scenes is different; generating a failure seed parameter according to the scene description information; the failure seed parameter is a parameter triggering the training scene; generating the training scene according to the failure seed parameter; the training scene is used for training the operation behavior of a person under test on the device in the training scene; the scene description information comprises target description sub-information of the device in multiple running stages, and the multiple running stages have a running sequence; the obtaining of the scene description information comprises: obtaining multiple target description sub-information; the multiple target description sub-information is respectively used for describing multiple sub-training scenes in the training scene, and the multiple target description sub-information is used for respectively generating failure seed parameters in the sub-training scenes corresponding to the multiple running stages; the generating of the training scene according to the failure seed parameter comprises: generating the sub-training scenes in the multiple running stages in sequence according to the failure seed parameters in the sub-training scenes corresponding to the multiple running stages according to the running sequence of the multiple running stages; the obtaining of the target description sub-information comprises: in a case where a condition vector input into multiple condition input boxes is received, obtaining the target description sub-information according to multiple condition vectors; the condition vector is used for representing basic information of the sub-training scene, and the multiple condition input boxes are associated with each other; a condition vector input into a previous condition input box is used for limiting the range of a condition vector input into a next condition input box.

2. The method of claim 1, wherein, the generating of the failure seed parameter according to the scene description information comprises: inputting the scene description information into a generator to obtain the failure seed parameter; the generator is obtained by training a training sample; the training sample comprises scene description information samples in different training scene samples, failure seed parameter samples corresponding to the scene description information samples, and association description information between the scene description information samples and the failure seed parameter samples; the association description information is used for describing the association relationship between the scene description information samples and the failure seed parameter samples; the generating of the training scene according to the failure seed parameter comprises: inputting the failure seed parameter into a renderer to generate the training scene; the renderer is used for generating audio and / or video generated by device running in the training scene.

3. The method of claim 1, wherein, each running stage corresponds to multiple description sub-information; the obtaining of the multiple target description sub-information comprises: obtaining the multiple description sub-information corresponding to each running stage; from the multiple description sub-information, the target description sub-information is screened out as the description information of the sub-training scene in the running stage; wherein, the sub-training scenes described by the description sub-information in the multiple running stages are combined to form a training scene, and the training scene is different from a presented historical training scene.

4. The method of claim 1, wherein, the obtaining of the scene description information comprises: screening, from scene description information of the multiple training scenes, scene description information of a training scene with a training probability less than a first preset probability and / or an accuracy less than a second preset probability; the training probability is used to represent a probability of exhibiting the training scene, and the accuracy is used to represent an accuracy of an operation behavior of the tested personnel in the training scene.

5. The method of claim 1, wherein, After the training scene is generated, the method further includes: obtaining operation behavior data of the tested personnel in the training scene; determining a target operation score of the tested personnel according to the operation behavior data; the target operation score is positively correlated with operation accuracy and / or timeliness of the tested personnel.

6. The method of claim 5, wherein, The operation behavior data includes the operation behavior and a time point at which the operation behavior is generated. The determining of the target operation score of the tested personnel according to the operation behavior data includes: obtaining a first operation score according to an accuracy of the operation behavior; the accuracy of the operation behavior is used to represent a proportion of correct operation behaviors in an operation behavior quantity of the operation behavior performed by the tested personnel, and the operation behavior quantity is a quantity of preset operation behaviors expected in the training scene; obtaining a second operation score according to a time deviation between the time point of the operation behavior and a preset time point; the preset time point is used to represent a time point at which the correct operation behavior occurs in the training scene; obtaining the target operation score according to the first operation score and / or the second operation score.

7. The method of claim 6, wherein, The operation behavior data further includes a target decision made by the tested personnel after the operation behavior is performed; After the training scene is generated, the method further includes: generating an analysis report according to a contribution degree of the operation behavior performed by the tested personnel before the target decision is made to the target decision; the contribution degree is used to represent a contribution of the operation behavior to promoting the tested personnel to make the target decision, and the analysis report is used to represent a cause analysis of the tested personnel to make the target decision.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The processor implements the method of any one of claims 1 to 7 when executing the computer program.

Images