Testing method and device based on unreal engine and electronic equipment

Abstract

The invention discloses a testing method and device based on an unreal engine and electronic equipment, and belongs to the technical field of computers. The method comprises the steps of obtaining a virtual scene graph constructed based on an unreal engine; under the condition that the virtual scene graph is a blank scene, loading a first target article associated with the first target task in the virtual scene graph; controlling the intelligent agent to execute a first target task for the first target article in the virtual scene graph; and generating a test report according to an execution result of executing the first target task by the agent. According to the method, the objects related to the specific task are dynamically loaded under the condition that the virtual scene graph is the blank scene, manual complex setting is not needed, the loading mode can be flexibly customized according to test requirements, the flexibility of test scene construction is improved, and further, the test efficiency is improved. The target task is executed in the virtual scene graph through the intelligent agent, and the test report is automatically generated, so that the operation in a real environment can be simulated, and the test efficiency is greatly improved.

Description

Technical Field

[0001] The present application belongs to the field of computer technology, and in particular, relates to a testing method, device and electronic equipment based on Unreal Engine. Background Art

[0002] With the rapid advancement of virtual reality technology, UE (Unreal Engine) has become the first choice for building high-precision simulation platforms with its powerful rendering, physical simulation and highly customizable capabilities. In the fields of robot navigation, game development, etc., UE can build a variety of simulation scenes that accurately simulate complex environments, which greatly facilitates algorithm and system testing.

[0003] In the related technologies, UE-based simulation test tools mostly focus on platform construction and scene rendering. When there is a test requirement, they rely on manual configuration to test a specific scene. When faced with multiple scenes with numerous objects and complex navigation logic, it is difficult to perform rapid testing. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes a test method, device and electronic device based on Unreal Engine to realize automated testing and improve test efficiency.

[0005] In a first aspect, the present application provides a testing method based on Unreal Engine, comprising: Get a virtual scene graph built based on Unreal Engine; When the virtual scene graph is a blank scene, loading a first target object associated with a first target task into the virtual scene graph; Controlling the intelligent agent to perform the first target task for the first target object in the virtual scene graph; A test report is generated according to the execution result of the agent performing the first target task.

[0006] According to the Unreal Engine-based testing method of the present application, a virtual scene graph constructed based on the Unreal Engine is obtained; when the virtual scene graph is a blank scene, a first target item associated with a first target task is loaded into the virtual scene graph; an intelligent agent is controlled to perform the first target task for the first target item in the virtual scene graph; and a test report is generated according to the execution result of the first target task performed by the intelligent agent. The embodiment of the present application obtains a virtual scene graph, and dynamically loads items related to a specific task when the virtual scene graph is a blank scene, without the need for manual complex settings. Furthermore, by having the intelligent agent perform the target task in the virtual scene graph and automatically generate a test report, it is possible to simulate operations in a real environment, greatly improving the efficiency of the test.

[0007] According to an embodiment of the present application, loading a first target object associated with a first target task in the virtual scene graph includes: Determine an item corresponding to the first target task from a task-item combination relationship table; Determine from the item table a first target item set of the same type as the items corresponding to the first target task; A first target object in the first target object set is loaded into the virtual scene graph.

[0008] In this embodiment, by determining the items corresponding to the first target task from the task-item combination relationship table, the loaded items can be highly correlated with the task, thereby improving the pertinence and effectiveness of the test, and the first target item set of the same type as the first target task is screened from the item table, and the same type of items can be efficiently batch tested, greatly improving the test efficiency. In addition, newly added items of the same type will also be automatically included in the test scope, further improving the scalability of the test.

[0009] According to one embodiment of the present application, the controlling agent performs the first target task for the first target object in the virtual scene graph, including: The intelligent agent is controlled to perform the first target task once for each first target object in the first target object set in the virtual scene graph.

[0010] In this embodiment, by testing each first target object in the first target object set one by one, the testing of different objects of the same type can be completed at one time, which greatly improves the testing efficiency.

[0011] According to one embodiment of the present application, the controlling agent performs the first target task for the first target object in the virtual scene graph, including: Acquire the execution action of the first target task from the action library of the Unreal Engine; The agent is controlled to perform the first target task on the first target object in the virtual scene using the execution action.

[0012] In this embodiment, by using the predefined action library in the Unreal Engine, the actions of the agent in performing tasks can be standardized, which helps to reduce the deviation of test results caused by inconsistent action execution. The action library allows developers to quickly adjust and configure the behavior of the agent according to different task requirements, so that the test can cover more scenarios and conditions.

[0013] According to an embodiment of the present application, generating a test report according to the execution result of the agent performing the first target task includes: Asserting the execution result of the agent performing the first target task; Generate a test report based on the assertion results.

[0014] In this embodiment, assertions are used to verify the correctness of the agent's task execution, and it is possible to automatically check whether the agent's behavior meets expectations. A test report is generated based on the assertion results, thereby achieving automatic recording and reporting of test results, which not only saves time in manually writing reports, but also improves the consistency and readability of reports.

[0015] According to one embodiment of the present application, the method further includes: Recording the execution process and the execution result of the first target task to obtain a process video; When the test report feedbacks that the execution result does not meet expectations, the test problem is located according to the process video and / or the assertion result.

[0016] In this embodiment, the execution process video is obtained by screen recording, which provides a visual record of the test process, allowing developers to intuitively observe the behavior of the intelligent agent in the virtual scene. When the test report shows that the execution result does not meet expectations, the process video or assertion result can be used to accurately locate the test problem, greatly improving the efficiency and accuracy of problem diagnosis.

[0017] According to one embodiment of the present application, the method further includes: Determine a second target task and a second target item corresponding to the test question; Initializing the virtual scene graph to a blank scene, and loading the second target object into the virtual scene graph; Preprocessing the second target object in the virtual scene graph, wherein the preprocessing includes at least one of rotating an angle, adding, deleting, modifying attributes, and dragging; Controlling the intelligent agent to perform the second target task for the second target object in the virtual scene; A test report is generated according to the execution process and execution result of the agent in executing the second target task.

[0018] In this embodiment, by determining the second target task and the second target object corresponding to the test problem and loading these objects in the virtual scene graph, it is possible to reproduce the specific test problem and pre-process the second target object, such as rotating the angle, adding, deleting and modifying attributes, dragging and other operations, to simulate specific conditions that may cause the test problem, thereby accurately locating the root cause of the problem and providing a reliable basis for repair.

[0019] According to one embodiment of the present application, the method further includes: When the virtual scene graph is not a blank scene, obtaining item information in the virtual scene graph; Determine an executable task list from a task and item combination relationship table according to the item information; Controlling the intelligent agent to execute the tasks in the task list in the virtual scene; A test report is generated according to the execution results of the tasks in the task list by the agent.

[0020] In this embodiment, by obtaining object information in a non-blank virtual scene graph and determining an executable task list based on this information, all objects in the virtual scene graph can be tested, thereby improving the automation and intelligence level of the test. Furthermore, the task list is generated based on the object information in the virtual scene graph, so that the test is not restricted by the scene, and can efficiently enable rapid verification of preset tasks in multiple scenes and newly added scenes, thereby improving test efficiency and scalability.

[0021] In a second aspect, the present application provides a testing device based on Unreal Engine, comprising: Initialization module, used to obtain the virtual scene graph built based on Unreal Engine; A loading module, used for loading a first target object associated with a first target task into the virtual scene graph when the virtual scene graph is a blank scene; An execution module, used for controlling the agent to execute the first target task for the first target object in the virtual scene graph; A generation module is used to generate a test report based on the execution result of the agent performing the first target task.

[0022] According to the Unreal Engine-based testing device of the present application, a virtual scene graph constructed based on the Unreal Engine is obtained; when the virtual scene graph is a blank scene, a first target item associated with a first target task is loaded into the virtual scene graph; an intelligent agent is controlled to perform the first target task for the first target item in the virtual scene graph; and a test report is generated according to the execution result of the first target task performed by the intelligent agent. The embodiment of the present application obtains a virtual scene graph, and dynamically loads items related to a specific task when the virtual scene graph is a blank scene, without the need for manual complex settings. Furthermore, by having the intelligent agent perform the target task in the virtual scene graph and automatically generate a test report, it is possible to simulate operations in a real environment, greatly improving the efficiency of the test.

[0023] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the Unreal Engine-based testing method as described in the first aspect above is implemented.

[0024] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the Unreal Engine-based testing method as described in the first aspect above.

[0025] In a fifth aspect, the present application provides a chip, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the Unreal Engine-based testing method as described in the first aspect above.

[0026] In a sixth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the Unreal Engine-based testing method as described in the first aspect above.

[0027] The above one or more technical solutions in the embodiments of the present application have at least one of the following technical effects: According to the Unreal Engine-based testing method of the present application, a virtual scene graph constructed based on the Unreal Engine is obtained; when the virtual scene graph is a blank scene, a first target item associated with a first target task is loaded into the virtual scene graph; an intelligent agent is controlled to perform the first target task for the first target item in the virtual scene graph; and a test report is generated according to the execution result of the first target task performed by the intelligent agent. The embodiment of the present application obtains a virtual scene graph, and dynamically loads items related to a specific task when the virtual scene graph is a blank scene, without the need for manual complex settings. Furthermore, by having the intelligent agent perform the target task in the virtual scene graph and automatically generate a test report, it is possible to simulate operations in a real environment, greatly improving the efficiency of the test.

[0028] Furthermore, in some embodiments, by determining the items corresponding to the first target task from the task-item combination relationship table, the loaded items can be highly correlated with the task, thereby improving the pertinence and effectiveness of the test, and by screening out the first target item set of the same type as the first target task from the item table, batch testing of the same type of items can be performed efficiently, greatly improving the test efficiency. In addition, newly added items of the same type will also be automatically included in the test scope, further improving the scalability of the test.

[0029] Furthermore, in some embodiments, by testing each first target item in the first target item set one by one, the testing of different items of the same type can be completed at one time, which greatly improves the testing efficiency.

[0030] Furthermore, in some embodiments, by using the predefined action library in the Unreal Engine, the actions of the agent in performing tasks can be standardized, which helps to reduce the deviation of test results caused by inconsistent action execution, and the action library allows developers to quickly adjust and configure the behavior of the agent according to different task requirements, so that the test can cover more scenarios and conditions.

[0031] Furthermore, in some embodiments, assertions are used to verify the correctness of the agent's task execution, which can automatically check whether the agent's behavior meets expectations, generate a test report based on the assertion results, and realize automatic recording and reporting of test results, which not only saves the time of manual report writing, but also improves the consistency and readability of the report.

[0032] Furthermore, in some embodiments, the execution process video is obtained by screen recording, which provides a visual record of the test process, allowing developers to intuitively observe the behavior of the intelligent agent in the virtual scene. When the test report shows that the execution result does not meet expectations, the process video or assertion result can be used to accurately locate the test problem, greatly improving the efficiency and accuracy of problem diagnosis.

[0033] Furthermore, in some embodiments, by determining the second target task and the second target object corresponding to the test problem and loading these objects in the virtual scene graph, it is possible to reproduce the specific test problem and pre-process the second target object, such as rotating the angle, adding, deleting and modifying attributes, dragging and other operations, to simulate specific conditions that may cause the test problem, thereby accurately locating the root cause of the problem and providing a reliable basis for repair.

[0034] Furthermore, in some embodiments, by obtaining object information in a non-blank virtual scene graph and determining an executable task list based on this information, all objects in the virtual scene graph can be tested, thereby improving the automation and intelligence level of the test. Furthermore, the task list is generated based on the object information in the virtual scene graph, so that the test is not restricted by the scene, and can efficiently enable rapid verification of preset tasks in multiple scenes and newly added scenes, thereby improving test efficiency and scalability.

[0035] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0037] Figure 1 It is a flowchart of a testing method based on Unreal Engine provided in an embodiment of the present application; Figure 2 is a schematic diagram of a virtual scene graph provided in an embodiment of the present application; Figure 3 is a schematic diagram of a blank scene provided in an embodiment of the present application; Figure 4 This is one of the schematic diagrams of the embodiment of the present application after loading items in a blank scene; Figure 5 This is the second schematic diagram of the embodiment of the present application after loading items in a blank scene; Figure 6 is a schematic diagram of a scenario example provided in an embodiment of the present application; Figure 7 is a schematic diagram of the structure of a test device based on Unreal Engine provided in an embodiment of the present application; Figure 8 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0039] The terms "first", "second", etc. in this application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are usually a class, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" in this application represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0040] The following is a detailed description of the Unreal Engine-based testing method, device, and electronic device provided in the embodiments of the present application through specific embodiments and their application scenarios in conjunction with the accompanying drawings.

[0041] Among them, the testing method based on Unreal Engine can be applied to the terminal, and can be specifically executed by hardware or software in the terminal.

[0042] The terminal includes, but is not limited to, a portable communication device such as a mobile phone or a tablet computer having a touch-sensitive surface (e.g., a touch screen display and / or a touch pad). It should also be understood that in some embodiments, the terminal may not be a portable communication device, but a desktop computer having a touch-sensitive surface (e.g., a touch screen display and / or a touch pad).

[0043] In the following various embodiments, a terminal including a display and a touch-sensitive surface is described. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, a mouse and a joystick.

[0044] The Unreal Engine-based testing method provided in the embodiment of the present application may be executed by an electronic device or a functional module or functional entity in the electronic device that can implement the Unreal Engine-based testing method. The electronic devices mentioned in the embodiment of the present application include but are not limited to mobile phones, tablet computers, computers, cameras, and wearable devices. The Unreal Engine-based testing method provided in the embodiment of the present application is described below using an electronic device as an example of an execution subject.

[0045] like Figure 1 As shown, the Unreal Engine-based testing method includes: step 110 , step 120 , step 130 and step 140 .

[0046] Step 110: Obtain a virtual scene graph constructed based on the Unreal Engine.

[0047] In the embodiment of the present application, Unreal Engine is a powerful game engine. With the continuous evolution of technology, it has now become a cross-platform tool widely used in multiple fields such as game development, film production, architectural visualization, virtual reality (VR), augmented reality (AR), etc. The virtual scene graph is a visual representation of the scene created in the virtual environment based on the Unreal Engine. The virtual scene graph can present various elements of the virtual world, such as terrain, buildings, light sources, objects and characters. The virtual scene graph can be used not only for game and film production, but also for various application scenarios such as architectural design, training simulation, product display, etc.

[0048] In the embodiment of the present application, the virtual scene graph may be a virtual scene established based on the test task. For example, in the test task of embodied intelligent scene, different virtual scene graphs may be established in advance, such as Figure 2 As shown, these virtual scene graphs may include household layouts, rooms, and items in the rooms, such as tables, cups, toys, beds, etc. The virtual scene graph may also include an agent, which may be a role created in the virtual scene graph, such as a specific character.

[0049] Step 120: When the virtual scene graph is a blank scene, load a first target object associated with the first target task into the virtual scene graph.

[0050] In the embodiment of the present application, in some test requirements, it may be necessary to only test the effect of the agent performing actions on the object, so a blank scene can be created, such as Figure 3 As shown, this blank scene includes no pre-built objects and assets.

[0051] In an embodiment of the present application, the virtual scene graph can be analyzed. If the virtual scene graph is identified as a blank scene, the first target task to be executed can be obtained. The first target task can be pre-set. For example, a task list can be pre-set, and the tasks in the task list are the first target tasks. In an example, the first target task can be putting a cup on a table, putting a toy on a sofa, etc. The embodiment of the present application does not limit the tasks.

[0052] In the embodiments of the present application, tasks are usually associated with objects. For example, if the task is to put a cup on a table, then the task is associated with the object cup and the object table. In this way, the first target object associated with the first target task can be identified based on the content of the first target task, wherein the first target object can be one or more.

[0053] After the first target object is determined, the Unreal Engine can be called to load the first target object into the virtual scene graph. For example, model files of various objects can be pre-set, and after the first target object is determined, the first target object in the model file can be imported into the virtual scene graph through the import function, such as Figure 4 After the first target object is imported into the virtual scene graph, the state of the first target object, such as position, rotation, scaling, etc., can be set to make the first target object meet the task requirements and scene layout.

[0054] Step 130: Control the intelligent agent to perform a first target task for a first target object in the virtual scene graph.

[0055] In an embodiment of the present application, an agent is a character that can act or react autonomously in a virtual scene graph, for example, the agent can be a character in the virtual scene graph. In Unreal Engine, a new blueprint class can be created, and a suitable base class, such as Character or Pawn, can be selected to define basic properties for the agent, which may include movement speed, perception range, etc. The navigation system, perception system, etc. can be pre-set. The navigation system is used to guide the agent to move in the virtual environment in UE. The navigation system can be implemented by constructing a navigation mesh (NavMesh), which is a pre-calculated data structure used to store information about traversable areas in the environment, so that the character can plan paths and move based on this information. The perception components of Unreal Engine can be used to configure the senses of the agent, such as vision, hearing, etc., so that the agent can perceive the surrounding objects.

[0056] When the agent starts to perform the first target task, the walking route can be navigated according to the position of the first target object in the virtual scene graph. During the process of performing the first target task, the agent can walk according to the walking route and perform corresponding actions when perceiving the first target object. For example, if the first target task is to place a cup on the table, the agent first walks to a position where the cup can be picked up, and after picking up the cup, walks to the vicinity of the table and places the cup on the table.

[0057] Step 140: Generate a test report based on the execution result of the agent performing the first target task.

[0058] In the embodiment of the present application, during the process of the agent performing the task, relevant execution data can be collected, such as the start and end time of the task, the execution path, the selection of decision points, interaction events, etc. These data can be obtained through the log system or screen recording.

[0059] The behavior of the agent can be analyzed through the execution data of the process of the agent performing the task, such as whether the agent frequently collides with obstacles, or whether the behavior is reasonable when approaching the target.

[0060] It is possible to analyze whether the execution result of the first target task meets expectations. Specifically, the state of the objects in the virtual scene graph after the agent executes the first target task can be obtained, and whether the execution result meets expectations can be determined based on the state of the objects. For example, for the task "place a cup on the table", after the agent completes the execution, it is determined whether the cup is on the table based on the relative position relationship between the cup and the table. If the cup is on the table, it can be considered that the execution result meets expectations.

[0061] In the embodiment of the present application, a test report can be generated based on the analysis of the execution results and the analysis of the execution process.

[0062] According to the Unreal Engine-based testing method of the present application, a virtual scene graph constructed based on the Unreal Engine is obtained; when the virtual scene graph is a blank scene, a first target item associated with a first target task is loaded into the virtual scene graph; an intelligent agent is controlled to perform the first target task for the first target item in the virtual scene graph; and a test report is generated according to the execution result of the first target task performed by the intelligent agent. The embodiment of the present application obtains a virtual scene graph, and dynamically loads items related to a specific task when the virtual scene graph is a blank scene, without the need for manual complex settings. Furthermore, by having the intelligent agent perform the target task in the virtual scene graph and automatically generate a test report, it is possible to simulate operations in a real environment, greatly improving the efficiency of the test.

[0063] In some embodiments, loading a first target object associated with a first target task in a virtual scene graph includes: Determine the item corresponding to the first target task from the task-item combination relationship table; Determine from the item table a first target item set of the same type as the items corresponding to the first target task; A first target item in a first target item set is loaded into the virtual scene graph.

[0064] In this embodiment, a table, database or configuration file may be predefined, and the table, database or configuration file may include the combination relationship between tasks and objects. For example, if the task is "put the cup on the table", the task is associated with the object cup and the object table. The task-object combination relationship table may be a simple Excel table, a database query result or a data structure defined in the code. According to the identifier or name of the first target task, the task-object combination relationship table is queried to find the object identifier or name associated with the first target task.

[0065] In an example, the combination relationship between tasks and items is defined as follows: [Quench your thirst] drink_milk=cup,drinkContainer drink_water=cup,waterDispenser / coffeeMachine,table / coffeetable / diningTable [clean] pick_up_paper_ball=paperBall,wastebasket clean_obj=sponge / rag,dirt mop_floor=mop,dirt wash=faucet,stool The task is "quenching thirst", and the first task is defined as using a cup to drink milk from a container to quench thirst; the second task is using a cup to drink water from a container or coffee machine, and the container or coffee machine is placed on a table, coffee table or dining table. The first target items associated with the thirst quenching task may include cups, containers, milk, tables, coffee tables, dining tables, etc.

[0066] The task is "clean", and the first task is defined as picking up paper balls and putting them into the trash can to keep the environment clean; the second task is to use a sponge or rag to clean the dirt on the object; and the third task is to use a mop to clean the dirt on the ground. The first target items associated with the clean task can include paper balls, trash cans, sponges, rags, mops, etc.

[0067] In this embodiment, after determining the items corresponding to the first target task, the items of the same type as these items can be determined. Specifically, an item table can be pre-set, and the items in the item table are items that can be loaded into the virtual scene graph. For example, the items in the item table can be items contained in a pre-set object model file, and the items not contained in the object model cannot be automatically loaded into the virtual scene graph for the time being. A set of items of the same type can be filtered out from the item table as the first target item set based on the items corresponding to the first target task. For example, if the items corresponding to the first target task include tables and cups, then a set of items of the same type as the table can be filtered out from the item table, such as dining tables, coffee tables, tea tables, desks, etc., and a set of items of the same type as the cup can be filtered out from the item table, such as water cups, mouthwash cups, coffee cups, wine glasses, etc.

[0068] It should be noted that different tasks may correspond to different combinations of items, and the first target items include these combinations of items. For example, if the item combination corresponding to Task 1 of the clean task is paper ball + trash can, then the first target item may include a combination of paper ball + trash can, and paper ball may correspond to multiple items of the same type, and trash can may also correspond to multiple items of the same type, then the first target item set may include a set of items of the same type as paper ball and a set of items of the same type as trash can. For another example, if the item combination corresponding to Task 2 of the clean task is sponge or rag + stain, then the first target item may include a combination of sponge + stain or a combination of rag + stain, and the first target item set may include a set of items of the same type as sponge and a set of items of the same type as stain, or the first target item set may include a set of items of the same type as rag and a set of items of the same type as stain.

[0069] After the first target item set is determined, the first target item in the first target item set may be loaded into the virtual scene graph.

[0070] In some embodiments, the loading process may select one item from the first target item set and one item from another item set for loading, or may select one item from the first target item set and multiple items from another item set for loading, or may select multiple items from the first target item set and one item from another item set for loading. For example, for the task "Put the cup on the table", a combination of a cup + a coffee table may be loaded, a combination of a cup + a coffee table and a dining table may be loaded, or multiple combinations of cups + coffee tables may be loaded, such as Figure 5 The loading method can be flexibly customized according to test needs, further improving the flexibility of test scenario construction.

[0071] In this embodiment, by determining the items corresponding to the first target task from the task-item combination relationship table, the loaded items can be highly correlated with the task, thereby improving the pertinence and effectiveness of the test, and the first target item set of the same type as the first target task is screened from the item table, and the same type of items can be efficiently batch tested, greatly improving the test efficiency. In addition, newly added items of the same type will also be automatically included in the test scope, further improving the scalability of the test.

[0072] In some embodiments, controlling the agent to perform a first target task for a first target object in a virtual scene graph includes: The intelligent agent is controlled to perform a first target task once for each first target object in the first target object set in the virtual scene graph.

[0073] In this embodiment, each first target object in the first target object set may be traversed, and for each first target object, the first target task needs to be executed once.

[0074] For example, if the first target task is "sleeping", and the first target item set includes a large bed, a small bed, a sofa, a tatami, etc., then the agent will perform the sleeping task for one of the items in the first target item set. After the task is completed, the agent will then perform the sleeping task for another item in the first target item set until the first target item in the first target item set is traversed. If the first target task is "placing a cup on the table", and the first target items include two types of items, cups and tables, then the first target item set may include a set of cups and a set of tables. A cup may be selected from the set of cups, and a table may be selected from the set of tables. The agent may perform the first target task once for the cup and table selected this time. For the next first target task, the agent may select a different table and a different cup, the same table and a different cup, or a different table and the same cup, etc., and perform the next first target task until the combination in the set of cups and the set of tables is traversed.

[0075] In this embodiment, by testing each first target object in the first target object set one by one, the testing of different objects of the same type can be completed at one time, which greatly improves the testing efficiency.

[0076] In some embodiments, controlling the agent to perform a first target task for a first target object in a virtual scene graph includes: Obtain the execution action of the first target task from the action library of the Unreal Engine; The intelligent agent is controlled to perform a first target task on a first target object in a virtual scene using an execution action.

[0077] In this embodiment, the action library of the Unreal Engine may include various preset actions, which may be pre-configured actions that can be used for animation and behavior control of the intelligent body.

[0078] In this embodiment, actions and tasks can be associated and stored in an action library. After determining the first target task, the execution action corresponding to the first target task can be matched from the action library. According to the execution action, the intelligent agent can be controlled to execute the first target task.

[0079] In an example, the execution action of a task is defined as follows: Task: Pick up the item and put it into the target item def do_take_and_putin(self, take_item, object_item): takeController = self.scene_manager.GetObjectByName(take_item) objectController = self.scene_manager.GetObjectByName(object_item) takeLocation = takeController.GetPose().location placeLocation = objectController.GetPlaceLocation() self.ai_baby.PutDownSthToLocation(placeLocation, True, UERotation(0,0, 0, 1), True, RightHand, True) Among them, do_take_ and _putin represent the execution of pick-up and put-in actions, take_item represents the name of the item to be picked up, and object_item represents the name of the target item. takeController and objectController represent obtaining the corresponding controller object according to the item name through the scene manager self.scene_manager. takeLocation represents obtaining the location of the picked-up item, and placeLocation represents the placement location of the target item. PutDownSthToLocation represents placing the item picked up by the agent at the target location, specifying whether it needs to be placed accurately, the rotation method, whether it needs to be aligned, which hand to use, and whether it needs to be confirmed.

[0080] In this embodiment, by using the predefined action library in the Unreal Engine, the actions of the agent in performing tasks can be standardized, which helps to reduce the deviation of test results caused by inconsistent action execution. The action library allows developers to quickly adjust and configure the behavior of the agent according to different task requirements, so that the test can cover more scenarios and conditions.

[0081] In some embodiments, generating a test report according to the execution result of the agent performing the first target task includes: Assert the execution result of the agent in executing the first target task; Generate a test report based on the assertion results.

[0082] In this embodiment, the expected results of the agent's task execution can be predefined, such as whether the agent successfully picks up an object, whether the object is correctly placed in the target location, etc., and then assertions can be written in the test script to verify the actual results. For example, an assert statement can be used to check whether the agent has reached a predetermined location or successfully picked up an object. In this embodiment, the assertion results can be collected and a test report can be generated based on the assertion results.

[0083] In this embodiment, assertions are used to verify the correctness of the agent's task execution, and it is possible to automatically check whether the agent's behavior meets expectations. A test report is generated based on the assertion results, thereby achieving automatic recording and reporting of test results, which not only saves time in manually writing reports, but also improves the consistency and readability of reports.

[0084] In some embodiments, the method further comprises: Recording the execution process and results of the first target task to obtain a process video; If the test report feedback shows that the execution results do not meet expectations, locate the test problems based on the process video and / or assertion results.

[0085] In this embodiment, a screen recording tool, screen recording software or the built-in screen recording function of the Unreal Engine can be used to record the process of the agent performing the task, including the walking process, behavioral performance, execution results, etc. of the agent. Specifically, the resolution, frame rate, and encoding format of the screen recording can be set so that the quality of the recorded video meets the analysis requirements and the file size is controllable. Start the screen recording tool before the agent starts to perform the task so that the entire process from the start to the end of the task is recorded. After the task is completed, stop the screen recording and save the video file.

[0086] In this embodiment, the test report can be analyzed. When the test report indicates that the execution result does not meet expectations, the specific information in the test report can be viewed, including failed assertions and error messages, and the screen recording video can be opened to compare the problem description in the test report, and the process of the agent performing the task can be reviewed to observe the behavior of the agent. Specifically, the behavior of the agent can be observed through the video, especially the key links such as the decision point, interactive actions and task results of the agent, and the specific link where the test problem occurs can be located in combination with the assertion results.

[0087] In this embodiment, the execution process video is obtained by screen recording, which provides a visual record of the test process, allowing developers to intuitively observe the behavior of the intelligent agent in the virtual scene. When the test report shows that the execution result does not meet expectations, the process video or assertion result can be used to accurately locate the test problem, greatly improving the efficiency and accuracy of problem diagnosis.

[0088] In some embodiments, the method further comprises: Determine the second target task and the second target item corresponding to the test question; Initialize the virtual scene graph to a blank scene, and load the second target object into the virtual scene graph; Preprocessing the second target object in the virtual scene graph, the preprocessing comprising at least one of rotating an angle, adding, deleting, modifying attributes, and dragging; Controlling the intelligent agent to perform a second target task for a second target object in the virtual scene; A test report is generated according to the execution process and execution results of the agent in executing the second target task.

[0089] In this embodiment, after locating the test problem, the second target task and the second target object corresponding to the test problem can be determined. For example, if the test problem is that the cup is not successfully placed on the table, then the second target task can be determined to be "place the cup on the table", and the second target object can be the cup and the table.

[0090] In this embodiment, a virtual scene graph including a blank scene can be initialized in the Unreal Engine, and the second target object can be loaded in the virtual scene graph. In order to analyze the cause of the test problem, the second target object can be preprocessed in the virtual scene graph, and the preprocessing can include rotation angle, adding, deleting, modifying attributes, dragging, etc. For example, the second target object can be rotated 30°, the attributes of the second target object can be added, such as lowering the height of the second target object, dragging the second target object, adjusting the position of the second target object, etc.

[0091] After preprocessing the second target object, the agent is controlled to perform the second target task for the second target object in the virtual scene, and a test report is generated based on the execution process and results of the agent's execution of the second target task. Finally, the root cause of the problem can be analyzed based on the content of this test report. For example, if the preprocessing is to rotate the second target object by 30°, and if the execution result of the task is shown in this test report to be in line with expectations, then the root cause of the problem may be that there is a problem with the placement angle of the second target object.

[0092] In this embodiment, by determining the second target task and the second target object corresponding to the test problem and loading these objects in the virtual scene graph, it is possible to reproduce the specific test problem and pre-process the second target object, such as rotating the angle, adding, deleting and modifying attributes, dragging and other operations, to simulate specific conditions that may cause the test problem, thereby accurately locating the root cause of the problem and providing a reliable basis for repair.

[0093] In some embodiments, the method further comprises: When the virtual scene graph is not a blank scene, obtaining item information in the virtual scene graph; Determine an executable task list from a task and item combination relationship table based on item information; Control the agent to perform tasks in the task list in the virtual scene; Generate a test report based on the execution results of the tasks in the agent's task list.

[0094] In this embodiment, when it is detected that the virtual scene graph is not a blank scene, Figure 2 As shown, the objects existing in the current virtual scene graph can be identified. For example, the object information of each object, such as the type, location, state, and attributes of the object, can be collected by scanning the scene or accessing the object database of the scene. Then, the executable task list is matched from the combination relationship table of tasks and objects in combination with the object information such as the type, location, state, and attributes of the object. The task list includes one or more tasks that can be executed by the agent. For example, the objects existing in the current virtual scene graph include a water cup. The attribute of the water cup is that the remaining liquid volume in the cup is 50 ml. Then, the task "pick up the water cup and drink water" can be matched. If the attribute of the water cup is that the remaining liquid volume in the cup is 0 ml, then the task "pick up the cup and drink water from the water dispenser" can be matched.

[0095] In this embodiment, the control agent executes the task process summary in the task list in the virtual scene, and can match the execution action corresponding to the current task from the action library of the Unreal Engine, so that the agent can execute the current task according to the execution action. After the current task is completed, the unexecuted task can be selected from the task list for execution.

[0096] In some embodiments, during the execution of the task, the focus is on whether the navigation path is reachable, that is, the navigation is from target A to target B, the starting point is at the target A position, and the agent should be at the target B position after completing the task. During the execution process, attention is paid to whether the navigation path is reasonable and whether there is a serious collision or obstruction with objects on the way.

[0097] In some embodiments, a screen recording tool, screen recording software, or the built-in screen recording function of the Unreal Engine can also be used to record the process of the agent performing the task, including the agent's walking process, behavioral performance, execution results, etc. Start the screen recording tool before the agent starts to perform the task, so that the entire process from the beginning to the end of the task is recorded. After the task is completed, stop the screen recording and save the video file.

[0098] After all the tasks in the task list are completed, a test report can be generated based on the execution results of each task. The test report can be analyzed. When the test report indicates that the execution result does not meet expectations, you can view the specific information in the test report, including failed assertions and error messages, open the screen recording video, and compare the problem description in the test report to review the process of the agent executing the task and observe the agent's behavior. Specifically, the agent's behavior can be observed through the video, especially the key links such as the agent's decision points, interactive actions, and task results, and combined with the assertion results, the specific link where the test problem occurs can be located. After determining the task corresponding to the test problem, the execution process of the agent can be reproduced according to the task and the corresponding item, and the root cause of the problem can be analyzed.

[0099] In this embodiment, by obtaining object information in a non-blank virtual scene graph and determining an executable task list based on this information, all objects in the virtual scene graph can be tested, thereby improving the automation and intelligence level of the test. Furthermore, the task list is generated based on the object information in the virtual scene graph, so that the test is not restricted by the scene, and can efficiently enable rapid verification of preset tasks in multiple scenes and newly added scenes, thereby improving test efficiency and scalability.

[0100] The following is a scenario example to illustrate the Unreal Engine-based testing method of the present application embodiment. Figure 6 As shown, in this scenario example, more than 100 virtual scene graphs are included, and traversal tests need to be performed on more than 100 virtual scene graphs. A blank map and item lists take_list and target_list can be loaded. After starting the test, the initialization step is entered, including starting the UE, initializing the virtual scene graph, and awakening the agent. If the currently traversed virtual scene graph is a blank scene, the item + action test branch can be performed, that is, the tested item is loaded in the blank scene. Specifically, the items take_items and target_items can be loaded through the BP_DebugLoadObjecActort function. Scenario 1 is used to test the effect of performing actions on items. In scenario 1, after loading is completed, the scene parsing step is entered. In this step, the items in the virtual scene graph can be analyzed, the item attributes can be verified, and the executable tasks can be determined. The executable tasks are matched from the combination relationship between the tasks and the items, and then the task execution step is entered. During the task execution process, the execution action corresponding to the task can be matched from the action library. The agent executes the task according to the execution action, and finally a test report is generated according to the execution result of the task.

[0101] In this scenario example, scenario two is used to quickly verify and reproduce test problems. In scenario two, objects can be preprocessed, such as horizontally rotating, adding, deleting, and modifying objects. After the preprocessing is completed, the scenario parsing step is entered.

[0102] In this scenario example, if the currently traversed virtual scene graph is not a blank scene, you can enter the multi-scenario test branch, identify the objects in the virtual scene graph, and enter the scene parsing step. In this step, you can analyze the objects in the virtual scene graph, verify the attributes of the objects, and determine the tasks that can be executed. Match the executable tasks from the combination relationship between the tasks and the objects, and then enter the task execution step. During the task execution process, you can match the execution action corresponding to the task from the action library, and the intelligent agent executes the task according to the execution action. Finally, a test report is generated based on the execution results of the task.

[0103] In this scenario example, if a new scenario needs to be added in a multi-scenario test, the name of the new virtual scene can be configured so that the newly added virtual scene graph can be automatically included in the test cycle. This method not only simplifies the test process, but also realizes efficient reuse of the test process, so that each time a new scenario is added, there is no need to repeatedly write test code, and the full set of tests can be automatically executed by simply updating the scenario configuration, which greatly improves the flexibility and efficiency of the test.

[0104] The Unreal Engine-based test method provided in the embodiment of the present application can be executed by a test device based on the Unreal Engine. In the embodiment of the present application, the Unreal Engine-based test device executing the Unreal Engine-based test method is taken as an example to illustrate the Unreal Engine-based test device provided in the embodiment of the present application.

[0105] The embodiment of the present application also provides a testing device based on Unreal Engine.

[0106] like Figure 7 As shown, the Unreal Engine-based test device includes: Initialization module 710, used to obtain a virtual scene graph built based on Unreal Engine; A loading module 720, configured to load a first target object associated with a first target task into the virtual scene graph when the virtual scene graph is a blank scene; An execution module 730 is used to control the agent to execute a first target task for a first target object in a virtual scene graph; The generating module 740 is used to generate a test report according to the execution result of the agent performing the first target task.

[0107] According to the Unreal Engine-based testing device of the present application, a virtual scene graph constructed based on the Unreal Engine is obtained; when the virtual scene graph is a blank scene, a first target item associated with a first target task is loaded into the virtual scene graph; an intelligent agent is controlled to perform the first target task for the first target item in the virtual scene graph; and a test report is generated according to the execution result of the first target task performed by the intelligent agent. The embodiment of the present application obtains a virtual scene graph, and dynamically loads items related to a specific task when the virtual scene graph is a blank scene, without the need for manual complex settings. Furthermore, by having the intelligent agent perform the target task in the virtual scene graph and automatically generate a test report, it is possible to simulate operations in a real environment, greatly improving the efficiency of the test.

[0108] In some embodiments, the loading module 720 is further configured to: Determine the item corresponding to the first target task from the task-item combination relationship table; Determine from the item table a first target item set of the same type as the items corresponding to the first target task; A first target item in a first target item set is loaded into the virtual scene graph.

[0109] In this embodiment, by determining the items corresponding to the first target task from the task-item combination relationship table, the loaded items can be highly correlated with the task, thereby improving the pertinence and effectiveness of the test, and the first target item set of the same type as the first target task is screened from the item table, and the same type of items can be efficiently batch tested, greatly improving the test efficiency. In addition, newly added items of the same type will also be automatically included in the test scope, further improving the scalability of the test.

[0110] In some embodiments, the loading module 720 is further configured to: The intelligent agent is controlled to perform a first target task once for each first target object in the first target object set in the virtual scene graph.

[0111] In some embodiments, the execution module 730 is further configured to: Obtain the execution action of the first target task from the action library of the Unreal Engine; The intelligent agent is controlled to perform a first target task on a first target object in a virtual scene using an execution action.

[0112] In some embodiments, the generating module 740 is further configured to: Assert the execution result of the agent in executing the first target task; Generate a test report based on the assertion results.

[0113] In some embodiments, the generating module 740 is further configured to: Recording the execution process and results of the first target task to obtain a process video; If the test report feedback shows that the execution results do not meet expectations, locate the test problems based on the process video and / or assertion results.

[0114] In some embodiments, the generating module 740 is further configured to: Determine the second target task and the second target item corresponding to the test question; Initialize the virtual scene graph to a blank scene, and load the second target object into the virtual scene graph; Preprocessing the second target object in the virtual scene graph, the preprocessing comprising at least one of rotating an angle, adding, deleting, modifying attributes, and dragging; Controlling the intelligent agent to perform a second target task for a second target object in the virtual scene; A test report is generated according to the execution process and execution results of the agent in executing the second target task.

[0115] In some embodiments, the generating module 740 is further configured to: When the virtual scene graph is not a blank scene, obtaining item information in the virtual scene graph; Determine an executable task list from a task and item combination relationship table based on item information; Control the agent to perform tasks in the task list in the virtual scene; Generate a test report based on the execution results of the tasks in the agent's task list.

[0116] The test device based on Unreal Engine in the embodiment of the present application can be an electronic device, or a component in an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal, or other devices other than a terminal. Exemplarily, the electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, a car electronic device, a mobile Internet device (Mobile Internet Device, MID), an augmented reality (augmented reality, AR) / virtual reality (virtual reality, VR) device, a robot, a wearable device, an ultra-mobile personal computer (ultra-mobilepersonal computer, UMPC), a netbook or a personal digital assistant (personal digital assistant, PDA), etc., and can also be a server, a network attached storage (Network Attached Storage, NAS), a personal computer (personal computer, PC), a television (television, TV), a teller machine or a self-service machine, etc., and the embodiment of the present application is not specifically limited.

[0117] The Unreal Engine-based test device in the embodiment of the present application may be a device having an operating system. The operating system may be a Microsoft (Windows) operating system, an Android (Android) operating system, an IOS operating system, or other possible operating systems, which are not specifically limited in the embodiment of the present application.

[0118] In some embodiments, Figure 8 As shown, an embodiment of the present application further provides an electronic device 800, including a processor 801, a memory 802, and a computer program stored in the memory 802 and executable on the processor 801. When the program is executed by the processor 801, the various processes of the above-mentioned Unreal Engine-based testing method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0119] It should be noted that the electronic devices in the embodiments of the present application include the above-mentioned mobile electronic devices and non-mobile electronic devices.

[0120] The embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the above-mentioned Unreal Engine-based testing method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0121] The processor is the processor in the electronic device in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0122] An embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned Unreal Engine-based testing method when executed by a processor.

[0123] The processor is the processor in the electronic device in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.

[0124] The embodiment of the present application further provides a chip, which includes a processor and a communication interface, the communication interface and the processor are coupled, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned Unreal Engine-based testing method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0125] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0126] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0127] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0128] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose and scope of protection of the present application, all of which are within the protection of the present application.

[0129] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0130] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present application, and that the scope of the present application is defined by the claims and their equivalents.

Claims

1. A testing method based on Unreal Engine, characterized in that: include: Get a virtual scene graph built based on Unreal Engine; When the virtual scene graph is a blank scene, loading a first target object associated with a first target task into the virtual scene graph; Controlling the intelligent agent to perform the first target task for the first target object in the virtual scene graph; A test report is generated according to the execution result of the agent performing the first target task.

2. The method according to claim 1, characterized in that The step of loading a first target object associated with a first target task in the virtual scene graph comprises: Determine an item corresponding to the first target task from a task-item combination relationship table; Determine from the item table a first target item set of the same type as the items corresponding to the first target task; A first target object in the first target object set is loaded into the virtual scene graph.

3. The method according to claim 2, characterized in that The controlling agent performs the first target task for the first target object in the virtual scene graph, including: The intelligent agent is controlled to perform the first target task once for each first target object in the first target object set in the virtual scene graph.

4. The method according to claim 1, characterized in that The controlling agent performs the first target task for the first target object in the virtual scene graph, including: Acquire the execution action of the first target task from the action library of the Unreal Engine; The agent is controlled to perform the first target task on the first target object in the virtual scene using the execution action.

5. The method according to claim 1, characterized in that The generating a test report according to the execution result of the agent performing the first target task includes: Asserting the execution result of the agent performing the first target task; Generate a test report based on the assertion results.

6. The method according to claim 1, characterized in that The method further comprises: Recording the execution process and the execution result of the first target task to obtain a process video; When the test report feedbacks that the execution result does not meet expectations, the test problem is located according to the process video and / or the assertion result.

7. The method according to claim 6, characterized in that The method further comprises: Determine a second target task and a second target item corresponding to the test question; Initializing the virtual scene graph to a blank scene, and loading the second target object into the virtual scene graph; Preprocessing the second target object in the virtual scene graph, wherein the preprocessing includes at least one of rotating an angle, adding, deleting, modifying attributes, and dragging; Controlling the intelligent agent to perform the second target task for the second target object in the virtual scene; A test report is generated according to the execution process and execution result of the agent in executing the second target task.

8. The method according to claim 1, characterized in that The method further comprises: When the virtual scene graph is not a blank scene, obtaining item information in the virtual scene graph; Determine an executable task list from a task and item combination relationship table according to the item information; Controlling the intelligent agent to execute the tasks in the task list in the virtual scene; A test report is generated according to the execution results of the tasks in the task list by the agent.

9. A test device based on Unreal Engine, characterized in that: include: Initialization module, used to obtain the virtual scene graph built based on Unreal Engine; A loading module, used for loading a first target object associated with a first target task into the virtual scene graph when the virtual scene graph is a blank scene; An execution module, used for controlling the agent to execute the first target task for the first target object in the virtual scene graph; A generation module is used to generate a test report based on the execution result of the agent performing the first target task.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 8 is implemented.

Images