A control method, device and apparatus for an object action, and a storage medium

Abstract

Embodiments of the present application disclose a kind of object action control method, device, equipment and storage medium.The method can include: when it is predicted that the second object of the same object team with the first object has action demand on the action object of the first object based on the first action prediction model, it is predicted that the first object and the second object jointly act on action object, whether can act successfully;If it is determined that the object combination including the first object and the second object can act successfully on the action object according to the prediction result of the first action prediction model, then based on the second action prediction model, it is predicted that the second object has action demand on the action object, it is predicted that the first object and the second object jointly act on action object, whether can act successfully;When it is determined that the object combination can act successfully on the action object according to the prediction result of the second action prediction model, control object combination to act on the action object.The technical scheme of the embodiment of the present application can realize the cooperative action between objects.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of computer, and particularly relate to a method and device for controlling object action, and a storage medium. BACKGROUND

[0002] Nowadays, with the rapid development of various technologies, there are more and more application scenarios in which robots replace human beings to perform certain tasks. However, when performing tasks, robots with low intelligence basically act alone, which makes it difficult for them to complete tasks with high quality. SUMMARY

[0003] Embodiments of the present application provide a method and device for controlling object action, and a storage medium, which solve the problem that objects cannot act in coordination.

[0004] In a first aspect, embodiments of the present application provide a method for controlling object action, which can include:

[0005] obtaining a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to the same object team as the first object;

[0006] predicting whether the second object has an action demand for the action object based on the first action prediction model, and if so, predicting whether the action can be successful when the first object and the second object jointly perform the action on the action object;

[0007] if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully act on the action object, predicting whether the second object has an action demand for the action object based on the second action prediction model, and if so, predicting whether the action can be successful when the first object and the second object jointly perform the action on the action object;

[0008] when it is determined according to the prediction result of the second action prediction model that the object combination can successfully act on the action object, controlling the object combination to perform the action on the action object.

[0009] In a second aspect, embodiments of the present application also provide a device for controlling object action, which can include:

[0010] a model obtaining module configured to obtain a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to the same object team as the first object;

[0011] a first action prediction module configured to predict whether the second object has an action demand for the action object based on the first action prediction model, and if so, predict whether the action can be successful when the first object and the second object jointly perform the action on the action object;

[0012] a second action prediction module, configured to, if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully act on the action object, predict whether the second object has an action demand on the action object based on a second action prediction model, and if yes, predict whether the action can be successfully performed when the first object and the second object jointly act on the action object;

[0013] an action control module, configured to, when it is determined according to the prediction result of the second action prediction model that the object combination can successfully act on the action object, control the object combination to act on the action object.

[0014] In a third aspect, an embodiment of the present application further provides a control device for object action, which can include:

[0015] one or more processors; a memory, configured to store one or more programs;

[0016] when the one or more programs are executed by the one or more processors, the one or more processors implement the control method for object action provided by any embodiment of the present application.

[0017] In a fourth aspect, an embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the control method for object action provided by any embodiment of the present application.

[0018]

[0019] The technical scheme of the embodiment of the present application comprises the following steps: obtaining a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to the same object team as the first object; predicting whether the second object has an action demand for the action object based on the first action prediction model, and if yes, predicting whether the action can be successful when the first object and the second object jointly perform the action on the action object; if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully perform the action on the action object, that is, when it is determined that the first object performs the action on the action object, the second action prediction model can be used to further predict whether the second object has an action demand for the action object, and if yes, predicting whether the action can be successful when the first object and the second object jointly perform the action on the action object; and when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform the action on the action object, controlling the object combination to perform the action on the action object. The above technical scheme predicts the action of the object itself and / or the action of the object not determined which belongs to the same object team as the object itself by using the action prediction model of the object, and controls each object in the object combination to perform the action on the action object in a team cooperation manner when the object combination including the objects corresponding to the at least two action prediction models can successfully perform the action on the action object, thereby achieving the effect of cooperative action among objects. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 is a flow chart of a control method of an object action in the first embodiment of the present application;

[0021] Figure 2 is a flow chart of a control method of an object action in the second embodiment of the present application;

[0022] Figure 3 is a flow chart of a control method of an object action in the third embodiment of the present application;

[0023] Figure 4 is a schematic diagram of various lead points and various positions in the control method of the object action in the third embodiment of the present application;

[0024] Figure 5 is a structural block diagram of a control device of an object action in the fourth embodiment of the present application;

[0025] Figure 6 is a structural schematic diagram of a control device of an object action in the fifth embodiment of the present application. DETAILED DESCRIPTION

[0026] The application will be described in further detail below with reference to the drawings and embodiments. It is to be understood that the specific embodiments described herein are merely illustrative of the application and are not to be used as limitations thereof. In addition, it is to be understood that, for the purpose of description, only the parts related to the application are shown in the drawings rather than all the structures.

[0027] Before introducing the embodiments of the application, the application scenarios of the embodiments of the application are exemplarily described: For the application scenarios described in the background art that robots replace manual work to perform certain tasks, examples are as follows: adding robot roles to fight against real players in a fighting game, carrying objects by robots in an object carrying task, assembling objects by robots in an object assembling task, etc., which are not specifically limited herein. At present, each robot basically acts alone, such as fighting against real players alone, carrying objects alone, assembling objects alone, etc. However, in some cases, the robots acting alone are difficult to complete the tasks with high quality, such as a single robot not having enough intelligence to fight against real players, being difficult to carry heavy objects, being difficult to assemble complex objects, etc. At this time, the robots need to perform team cooperation to complete the tasks together. Therefore, how to perform team cooperation among robots is a technical problem to be solved.

[0028] It should be noted that, on the one hand, the above is described by taking robots as an example, and in actual application, it can also be unmanned aerial vehicles, unmanned vehicles, unmanned ships, etc. objects having certain automatic performance, which are not specifically limited herein. On the other hand, in order to more intuitively understand the following embodiments, the fighting game is taken as an example for illustration, but the fighting game is only one application scenario of the following embodiments, and is not a specific limitation on the application scenario.

[0029] Embodiment one

[0030] Figure 1 is a flowchart of a method for controlling object action provided in the first embodiment of the application. The present embodiment can be applied to the case of controlling object action by predicting whether to perform team cooperation among objects. The method can be performed by the object action control device provided by the embodiments of the application, which can be realized by software and / or hardware, and can be integrated on an object action control device, which can be various user terminals or servers.

[0031] Referring to Figure 1 , the method of the embodiments of the application specifically includes the following steps:

[0032] S110, obtaining a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to the same object team as the first object.

[0033] The first action prediction model can be a model of the first object for predicting actions of the first object and / or other objects whose actions are not determined; the action object can be an object for which the first object has an action demand, but whether the first object actually performs an action on the object is to be determined; the second object can be an object that belongs to the same object team as the first object, i.e., an object that can perform team collaboration with the first object; and the second action prediction model can be a model of the second object for predicting actions of the second object and / or other objects whose actions are not determined. That is, when predicting based on the first action prediction model, the first object is taken as the starting point for prediction; similarly, when predicting based on the second action prediction model, the second object is taken as the starting point for prediction. In other words, each object can take itself as a single thinking dimension to determine whether to perform team collaboration to jointly perform an action on the action object by predicting actions of other objects whose actions are not determined.

[0034] In S120, whether the second object has an action demand for the action object is predicted based on the first action prediction model, and if so, whether the action on the action object can be successfully performed when the first object and the second object jointly perform the action on the action object is predicted.

[0035] In S120, whether the second object has an action demand for the action object is predicted based on the first action prediction model, and if so, whether the action on the action object can be successfully performed when the first object and the second object jointly perform the action on the action object is predicted.

[0036] In S130, if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully perform an action on the action object, whether the second object has an action demand for the action object is predicted based on the second action prediction model, and if so, whether the action on the action object can be successfully performed when the first object and the second object jointly perform the action on the action object is predicted.

[0037] The object combination includes the first object and the second object, or in other words, the object combination can include the first object and the second object and other objects, which are not limited herein. If it is determined according to the prediction result of the first action prediction model that the object combination can successfully act on the action object, that is, if the action can be successfully performed when the objects in the object combination jointly act on the action object, it means that the first object has been determined to act on the action object, that is, the first object can be an action-determined object, and at this time, the action of the remaining objects except the first object can be predicted based on the second action prediction model. On this basis, after it is determined whether the action can be successfully performed based on the prediction result of the first action prediction model, the object action control method can further include: if yes, associating the action object with the first object; if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully act on the action object, the method can include: if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully act on the action object, and the association result between the action object and the first object is obtained. In other words, when the object combination can successfully act on the action object, the action object is associated with the first object. Since the information among the objects is shared, after the association result between the action object and the first object is obtained, the control device can predict the action of the remaining objects except the first object based on the second action prediction model.

[0038] When the action prediction is performed based on the second action prediction model, it can first predict whether the second object has an action demand for the action object, and if yes, it can predict whether the action can be successfully performed when the first object and the second object jointly act on the action object. On this basis, optionally, otherwise (that is, the second object does not have an action demand for the action object), the second action prediction model ends the prediction. Further optionally, if it is determined that the action cannot be successfully performed when the first object and the second object jointly act on the action object, the prediction of the remaining action-undetermined objects can be continued.

[0039] It should be noted that the first action prediction model and the second action prediction model can be the same or different action prediction models. Here, the two are distinguished in order to show which object is taken as the starting point for the prediction operation, and not to limit the specific content thereof.

[0040] S140, when it is determined according to the prediction result of the second action prediction model that the object combination can successfully act on the action object, controlling the object combination to act on the action object.

[0041] If the prediction result of the second action prediction model determines that the object combination can successfully act on the action object, it means that both the first object and the second object determine that the object combination can successfully act on the action object, and then the object combination can be controlled to act on the action object. In other words, if at least two objects in the object combination determine that the object combination can successfully act on the action object, the object combination is controlled to act. On this basis, optionally, in order to improve the probability of successful action, the object combination can be controlled to act when each object in the object combination determines that the object combination can successfully act on the action object.

[0042] It should be noted that the above technical solution takes a single object as a single thinking dimension, and determines whether to perform team cooperation through the information of each object. Specifically, an object can predict the action of an action-undetermined object belonging to the same object team as itself while assuming its own action, and determine its own action according to the prediction result. Thus, according to the determined action of some objects in the object team, it can be determined whether to perform team cooperation, thereby improving the intelligence of each object and enabling team cooperation among objects.

[0043] The technical solution of the embodiment of the application obtains a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to the same object team as the first object. Based on the first action prediction model, it is predicted whether the second object has an action demand for the action object. If yes, it is predicted whether the action can be successful when the first object and the second object jointly act on the action object. If the prediction result of the first action prediction model determines that the object combination including the first object and the second object can successfully act on the action object, that is, when the first object has determined to act on the action object, the second action prediction model can be used to further predict whether the second object has an action demand for the action object. If yes, it is predicted whether the action can be successful when the first object and the second object jointly act on the action object. If the prediction result of the second action prediction model determines that the object combination can successfully act on the action object, the object combination is controlled to act on the action object. The above technical solution predicts the action of the object itself and / or an action-undetermined object belonging to the same object team as the object through the action prediction model of the object. When at least two action prediction models predict that an object combination including objects corresponding to the at least two action prediction models can successfully act on an action object, the objects in the object combination are controlled to perform team cooperation to act on the action object, thereby achieving the effect of cooperative action among objects.

[0044] An optional technical solution, the control method of the object action can further include: determining whether there is an object that can be successfully acted on by the first object in each candidate object within a preset range of the first object, wherein the preset range includes a preset action range or a preset field of view range; if not, determining the action object from each candidate object. The preset action range can be a range in which the first object can act, the preset field of view range can be a range in which the first object can observe the remaining objects, and the candidate object can be an object within the preset range. Since the object that cooperates in the team is an object that cannot act successfully alone, when there is no object that can be successfully acted on by the first object in each candidate object, i.e., when the first object cannot act successfully on any candidate object alone, the action object that needs to act with the remaining objects can be determined from each candidate object to perform the subsequent prediction step.

[0045] In order to more intuitively understand the specific implementation process of each step, the following will be described by taking the attack process in a confrontation type game as an example. It should be noted that the terms of attack, offense, kill, etc. described in the following are optional solutions of the action, and are not specific limitations of the action. For example, taking the first object as a robot A (hereinafter referred to as A) as an example, whether there is a friendly teammate, an enemy target or an obstacle, etc. in the adjacent position adjacent to the current position of A is determined to determine whether A has a road to walk;

[0046] 1) If there is no road to walk, the most valuable action is performed, specifically, first attack the enemy target within the preset attack range that can be successfully killed, optionally, when the attack-defense of A > the blood of the enemy target, the enemy target can be successfully killed; if there is no such enemy target (i.e., an enemy target that can be independently killed), the enemy target that the friendly teammate has determined to attack within the preset attack range can be attacked, i.e., assisting the friendly teammate to attack; if there is also no such enemy target, the enemy target with the lowest blood within the preset attack range can be selected for attack. In actual application, optionally, the preset attack range can be a range formed by the adjacent positions adjacent to the current position of A.

[0047] 2) If there is a road to walk, all enemy targets within the preset field of view range are traversed, and the enemy targets that have been locked by the friendly teammate and can be killed by the attack in this round are excluded, on this basis, it is determined whether there is an enemy target that can kill the friendly teammate in this round, if so, the enemy target is determined as the attack target in this round. In actual application, optionally, the preset field of view range can be a range formed by positions within a preset distance threshold from the current position of A.

[0048] 3) If the enemy target that can be attacked is still not determined after the above steps, it is predicted whether to perform team cooperation. Specifically, for an enemy target X within the preset attack range of A, a teammate of A (assuming robot B) within the preset attack range of X is determined, and it is predicted whether B will attack X (specifically, by using the action prediction model of A) if A attacks X by using recursive steps 1), 2) and 3); if not, A gives up attacking X, otherwise it is determined whether B can be counterattacked by X in the process of attacking X; if yes, A gives up attacking X, otherwise it is determined by the prediction result of the action prediction model of A that A and B can jointly kill X and will not be counterattacked in the process of attacking, and X is taken as the attack target of A in the current round. In actual application, optionally, each enemy target within the preset attack range of A can be traversed according to the distance between each enemy target and A, and if the enemy target cannot be killed through team cooperation, another enemy target is determined. Further optionally, when the number of teammates of A within the preset attack range of X is at least two, each teammate of A can be predicted one by one.

[0049] 4) If it is determined through the above steps that the attack target of A is X, X is associated with A, so that the action prediction model of B will not predict A again when predicting. In actual application, optionally, the association of X and A can be realized by recording the current coordinates of X as the moving coordinates of A, so that the action prediction model of B can determine which enemy target A wants to attack through the moving coordinates recorded in A.

[0050] The above example can effectively exert the effects of team cooperation and maximum possession, thereby improving the game winning rate and increasing the interactivity and interest of the game.

[0051] In order to further understand the specific implementation process of each step, the following takes an article carrying task as an example for illustration. Illustratively, when it is determined that robot A cannot carry an article X by itself, it is assumed that A will carry X, and it is predicted based on the action prediction model of A whether robot B needs to carry X, and if yes, it is predicted whether A and B can jointly carry X successfully; if yes, X is associated with A; after obtaining the association result of X and A, it is predicted based on the action prediction model of B whether B needs to carry X, and if yes, it is predicted whether A and B can jointly carry X successfully; if yes, A and B jointly carry X, thereby achieving the effect of completing the carrying task of X through the team cooperation of A and B.

[0052] Embodiment Two

[0053] Figure 2is a flowchart of the object action control method provided in Embodiment Two of the present application. The present embodiment is optimized based on the technical solutions described above. In the present embodiment, optionally, after predicting whether the action can be successfully performed based on the first action prediction model, the object action control method can further include: if not, obtaining a third object belonging to the same object team as the first object, and predicting whether the third object has an action demand for the action object; if yes, predicting whether the action can be successfully performed when the first object, the second object, and the third object jointly perform the action on the action object; accordingly, if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully perform the action on the action object, it can include: if it is determined according to the prediction result of the first action prediction model that the object combination including the first object, the second object, and the third object can successfully perform the action on the action object; accordingly, after predicting whether the action can be successfully performed based on the second action prediction model, the object action control method can further include: if not, predicting whether the third object has an action demand for the action object; if yes, predicting whether the action can be successfully performed when the first object, the second object, and the third object jointly perform the action on the action object; accordingly, when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform the action on the action object, controlling the object combination to perform the action on the action object can include: when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform the action on the action object, predicting whether the third object has an action demand for the action object based on a third action prediction model of the third object, if yes, predicting whether the action can be successfully performed when the first object, the second object, and the third object jointly perform the action on the action object; and when it is determined according to the prediction result of the third action prediction model that the object combination can successfully perform the action on the action object, controlling the object combination to perform the action on the action object. Wherein, the same or corresponding terms are not described here again.

[0054] Referring to Figure 2 The method of the present embodiment can specifically include the following steps:

[0055] S210, obtaining a first action prediction model of a first object, an action object, and a second action prediction model of a second object belonging to the same object team as the first object.

[0056] S220, predicting whether the second object has an action demand for the action object based on the first action prediction model, and if yes, predicting whether the action can be successfully performed when the first object and the second object jointly perform the action on the action object.

[0057] On this basis, optionally, if not (i.e., the second object does not have an action demand for the action object), the remaining objects in the same object team whose actions are not determined are taken as the second object, and the execution of S220 is returned.

[0058] S230, if no, a third object belonging to the same object team as the first object is obtained, and whether the third object has an action demand for the action object is predicted based on the first action prediction model.

[0059] The process of predicting the action of the third object based on the first action prediction model is similar to the process of predicting the action of the second object, and will not be described here.

[0060] On this basis, optionally, if yes (i.e. the first object and the second object can successfully act on the action object), the first action prediction model ends the prediction, and the second action prediction model is used to predict whether the second object has an action demand for the action object, and if yes, whether the first object and the second object can successfully act on the action object.

[0061] S240, if yes, the first action prediction model is used to predict whether the first object, the second object and the third object can successfully act on the action object.

[0062] On this basis, optionally, if no (i.e. the first object, the second object and the third object cannot successfully act on the action object), the remaining objects in the object team whose actions are not determined can be taken as the third object, and S230 is executed again.

[0063] S250, if the object combination including the first object, the second object and the third object can successfully act on the action object according to the prediction result of the first action prediction model, the second action prediction model is used to predict whether the second object has an action demand for the action object, and if yes, whether the first object and the second object can successfully act on the action object.

[0064] The object combination is a combination including the first object, the second object and the third object, in other words, in addition to the first object, the second object and the third object, the object combination can also include the remaining objects in the object team, which is not limited here. Since the first object has been determined to act on the action object, the actions of the remaining objects except the first object can be predicted based on the second action prediction model, such as the actions of the second object and the third object in the object combination.

[0065] S260, if no, the second action prediction model is used to predict whether the third object has an action demand for the action object.

[0066] S270, if yes, predicting whether the action on the action object can be successfully performed based on the second action prediction model when the first object, the second object and the third object jointly perform the action on the action object.

[0067] S280, when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform the action on the action object, predicting whether the third object has an action demand for the action object based on a third action prediction model of the third object, and if yes, predicting whether the action on the action object can be successfully performed when the first object, the second object and the third object jointly perform the action on the action object.

[0068] The prediction process of the third action prediction model is similar to that of the second action prediction model, and is used to predict the action of the object in the object team whose action is not determined, which will not be described herein again.

[0069] S290, when it is determined according to the prediction result of the third action prediction model that the object combination can successfully perform the action on the action object, controlling the object combination to perform the action on the action object.

[0070] In order to better understand the specific implementation process of the above technical solution, the first object A, the second object B, the third object C and the action object X are taken as examples for illustration. For example, it is assumed that A will perform the action on X, and it is predicted based on the first action prediction model whether B wants to perform the action on X, and if yes, it is predicted whether the combined force of A and B can successfully perform the action on X; if not, it is predicted based on the first action prediction model whether C wants to perform the action on X, and if yes, it is predicted whether the combined force of A, B and C can successfully perform the action on X; if yes, A is associated with X. Further, since A has determined to perform the action on X, it is predicted based on the second action prediction model whether B wants to perform the action on X, and if yes, it is predicted whether the combined force of A and B can successfully perform the action on X; if not, it is predicted based on the second action prediction model whether C wants to perform the action on X, and if yes, it is predicted whether the combined force of A, B and C can successfully perform the action on X; if yes, B is associated with X. Further, since A and B have determined to perform the action on X, it is predicted based on the third action prediction model whether C wants to perform the action on X, and if yes, it is predicted whether the combined force of A, B and C can successfully perform the action on X; if yes, C is associated with X, and A, B and C jointly perform the action on X, thereby achieving the effect of team cooperation of A, B and C.

[0071] The technical scheme of the embodiment of the present application, when the prediction result of the first action prediction model determines that the first object and the second object cannot successfully perform the action on the action object, the action of the third object belonging to the same object team can be predicted again; similarly, in addition to predicting the action of the second object based on the second action prediction model, the action of the third object can also be predicted; similarly, the action of the third object can also be predicted based on the third action prediction model, thereby achieving the effect of accurately determining whether to perform team cooperation among multiple objects through the mutual cooperation of each step.

[0072] Embodiment three

[0073] Figure 3 is a flowchart of an object action control method provided in the third embodiment of the present application. The present embodiment is optimized based on the above technical schemes. In the present embodiment, optionally, when the prediction performed based on the first action prediction model and / or the second action prediction model determines whether the first object and the second object can successfully perform the action on the action object, it can include: predicting a second preceding point of the second object in the action on the action object from each preceding point corresponding to the action object, wherein the preceding point includes a position adjacent to the action position of the action object; predicting whether the first object and the second object can successfully perform the action on the action object when the first object is at the first preceding point and the second object is at the second preceding point, wherein the first preceding point includes a preceding point of the first object in the action on the action object. Wherein, the same or corresponding terms as in the above embodiments are not repeated here. It should be noted that the following will be written based on the execution of the above steps based on the first action prediction model, but it can also be executed based on the second action prediction model, or simultaneously based on the first action prediction model and the second action prediction model, which is not specifically limited here.

[0074] Referring to Figure 3 , the method of the present embodiment can specifically include the following steps:

[0075] S310, obtaining a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to the same object team as the first object.

[0076] S320, predicting whether the second object has an action demand on the action object based on the first action prediction model, and if so, predicting a second preceding point of the second object in the action on the action object from each preceding point corresponding to the action object, wherein the preceding point includes a position adjacent to the action position of the action object.

[0077] Wherein, the preceding point can be a position adjacent to the action position of the action object, and the second preceding point can be a preceding point of the second object in the action on the action object among the preceding points.

[0078] S330, predicting, based on the first action prediction model, whether the action on the action object can be successfully performed when the first object is at the first preceding point and the second object is at the second preceding point.

[0079] The first preceding point can be a preceding point in which the first object is located when performing the action on the action object, and in actual applications, the first preceding point and the second preceding point can be determined by taking the first object moving to the first preceding point and the second object moving to the second preceding point as the goal, and the first preceding point can be predicted by the first action prediction model. When predicting based on the first action prediction model, it can be specifically predicted whether the action on the action object can be successfully performed when the first object is at the first preceding point and the second object is at the second preceding point.

[0080] S340, if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully perform the action on the action object, predicting, based on the second action prediction model, whether the second object has an action demand on the action object, and if so, predicting whether the action on the action object can be successfully performed when the first object and the second object jointly perform the action.

[0081] In actual applications, since the first preceding point and the action of the first object have been determined, when predicting based on the second action prediction model, in addition to predicting whether the second object has an action demand on the action object and whether the action on the action object can be successfully performed when the first object and the second object jointly perform the action, the second preceding point and / or whether the action on the action object can be successfully performed when the first object is at the first preceding point and the second object is at the second preceding point can also be predicted, and the like, which are not limited herein.

[0082] S350, when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform the action on the action object, controlling the object combination to perform the action on the action object.

[0083] The technical scheme of the embodiment of the application achieves the effect of accurately determining whether to perform team cooperation between objects by predicting the first preceding point and the second preceding point, and further predicting whether the action on the action object can be successfully performed when the first object is at the first preceding point and the second object is at the second preceding point.

[0084] On this basis, an optional technical solution is that the moving distance between the first object and the first preceding point is less than the moving distance between the second object and the second preceding point, a position adjacent to the first preceding point and not adjacent to the action position is taken as the first position, and a position adjacent to the second preceding point and not adjacent to the action position is taken as the second position; the control object combination performs an action on the action object, which can include: controlling the first object to move towards the first preceding point, and controlling the second object to move towards the second preceding point; when the first object moves to any first position, it is determined whether the first object needs to control the action object to avoid; if so, an avoidance position is determined from each first position, and the first object is controlled to move towards the avoidance position; when the second object moves to any second position, the first object is controlled to move to the first preceding point and the second object is controlled to move to the second preceding point; when the first object reaches the first preceding point and the second object reaches the second preceding point, the object combination including the first object and the second object performs an action on the action object.

[0085] In the control of the first object to move towards the first lead point and the control of the second object to move towards the second lead point, since the moving distance between the first object and the first lead point is less than the moving distance between the second object and the second lead point, when the moving speed of the first object and the second object is the same, the first object moves to the first lead point first and the second object moves to the second lead point later. On this basis, the position adjacent to the first lead point and not adjacent to the action position is taken as the first position, when the first object moves to any first position, it will move to the first lead point again after moving once, but at this time the second object has not moved to the second lead point. Since the action object can act on the first object while the first object can act on the action object, in order to avoid the situation that the first object is damaged by the action of the action object on the first object, when the first object reaches any first position, it can be determined whether it is necessary to control the first object to avoid the action object, which can be determined according to whether the action object is likely to damage the first object after the first object moves to the first lead point. If yes (i.e. avoidance is needed), the avoidance position is determined from each first position, and the first object is controlled to move towards the avoidance position, which can be a first position that can make the first object quickly move to the first lead point and the action object cannot act on the first object. Further, the position adjacent to the second lead point and not adjacent to the action position is taken as the second position, i.e. the second position can be a position that can make the second object quickly move to the second lead point, when the second object moves to any second position, the first object is controlled to move to the first lead point and the second object is controlled to move to the second lead point, and when the first object reaches the first lead point and the second object reaches the second lead point, each object in the object group is controlled to act on the action object in team cooperation, so that the situation that the object reaching the lead point first is damaged by the action object before team cooperation can be avoided.

[0086] In order to better understand the above concepts of lead point, first lead point, second lead point, first position and avoidance position, for example, as shown in Figure 4 Each square represents a position, assuming that 0 is the action position, then 1 is the lead point; assuming that the bold 1 is the first lead point, then 2 is the first position; assuming that the inclined and underlined 1 is the second lead point, then 3 is the second position.

[0087] On the basis of any of the above technical solutions, the action object includes an attack object, that is, the action of each object in the object combination on the attack object is an attack, and the method for controlling the object action further includes: if it is determined according to the prediction result of the first action prediction model that the object combination cannot attack the attack object successfully, obtaining a candidate acquisition object within a preset visual field range of the first object, and determining whether there is a to-be-acquired object that needs to be acquired among the candidate acquisition objects according to the need degree and the acquisition distance of the first object to each candidate acquisition object; if yes, obtaining a pursuit object of the first object, and determining whether to control the first object to pursue the pursuit object or to acquire the to-be-acquired object according to the pursuit distance between the first object and the pursuit object, the acquisition distance between the first object and the to-be-acquired object, and a priority coefficient of the to-be-acquired object relative to the pursuit object.

[0088] The attack object can be an object that can be attacked by the first object without moving, and the pursuit object can be an object that can be attacked by the first object after moving. The candidate acquisition object can be an object that can be acquired within the preset visual field range. For each candidate acquisition object, the need degree can be the degree to which the first object needs the candidate acquisition object, and the acquisition distance can be the moving distance between the first object and the candidate acquisition object. According to the need degree and the acquisition distance, it can be determined whether the first object really needs to acquire the candidate acquisition object at the current time. If yes, the candidate acquisition object can be taken as the to-be-acquired object. In the case where there is a to-be-acquired object among the candidate acquisition objects, the first object needs to move to the vicinity of the to-be-acquired object to acquire the to-be-acquired object, and in the process of moving, the other objects can move to the vicinity of the to-be-acquired object before the first object and acquire the to-be-acquired object first, that is, the process of acquiring the to-be-acquired object is risky. Therefore, even if there is a to-be-acquired object, the pursuit object of the first object can be acquired first. Similarly, the first object needs to move to the vicinity of the pursuit object to attack the pursuit object, and in the process of moving, the other objects can move to the vicinity of the pursuit object before the first object and defeat the pursuit object first, that is, the process of pursuing the pursuit object is also risky. Therefore, according to the pursuit distance between the first object and the pursuit object, the acquisition distance between the first object and the to-be-acquired object, and the priority coefficient of the to-be-acquired object relative to the pursuit object, it can be determined which of the pursuit risk and the acquisition risk is smaller, and then it can be determined whether to control the first object to pursue the pursuit object or to acquire the to-be-acquired object. For example, the first object can be controlled to pursue the pursuit object when the pursuit distance is less than the acquisition distance multiplied by the priority coefficient, and otherwise, the first object can be controlled to acquire the to-be-acquired object, thereby effectively controlling the first object to take the most valuable action.

[0089] To more intuitively understand the specific implementation process of each step, the following continues to take the above example of the confrontation game as an example to illustrate it. Illustratively, after determining that the first object cannot successfully kill each enemy target within the preset attack range, the first object can be controlled to perform a treasure hunting operation (i.e., an operation of obtaining an item to be obtained, and here the search can be understood as obtaining). The above enemy target can also be referred to as an attack target. Specifically,

[0090] 5) Prioritize each candidate treasure hunting object according to the degree of need and the search distance of each candidate treasure hunting object within the preset field of view of A, determine the to-be-searched treasure from each candidate treasure hunting object with the highest priority, without affecting the timely entry into the safe area, and without being locked by the friendly team, wherein the blood volume of the object in the safe area will not decrease, which means that when searching for treasure outside the safe area, it is possible to end the life due to the blood volume dropping out; if the to-be-searched treasure is not found, jump to step 7), otherwise, perform step 6).

[0091] 6) Determine the pursuit target of A according to steps 7)-8), and then determine whether to pursue the pursuit target or search for the to-be-searched treasure according to the pursuit distance, the search distance, and the priority coefficient.

[0092] 7) Sort the enemy targets within the preset field of view of A in ascending order of blood volume, defense, and movement distance, the purpose is to preferentially pursue the enemy target with the least blood volume, the lowest defense, and the nearest movement distance, i.e., the enemy target that is easiest to pursue and can be defeated by limited pursuit;

[0093] 8) Determine which enemy target to be the pursuit target. Specifically, each enemy target is processed in turn;

[0094] Exclude the enemy targets outside the safe area when the safe area is large, because when the safe area is large, there is no need to pursue the enemy targets outside the safe area, which requires a cost of reducing blood volume. Exclude the enemy targets that have been locked by the friendly team and can be killed. For the enemy targets that are not excluded, the following operations are performed in order according to their priority:

[0095] Suppose A is chasing an enemy target X, the first advance point of A is predicted, and whether A can kill X independently when moving to the first advance point is predicted. If not, whether B will chase X is predicted, and if so, the second advance point of B is predicted, and whether A and B can kill X together when A moves to the first advance point and B moves to the second advance point is predicted. If A and B cannot kill X together, the action of C is predicted, and the prediction process is similar to B, which will not be described here. Taking the determination that A and B can kill X together as an example, A is controlled to move towards the first advance point, and B is controlled to move towards the second advance point; when A has moved to any first position and B has not moved to any second position, it is determined whether A needs to be controlled to evade X to avoid the situation that A, which moves to the first advance point first, is attacked by X when B has not moved to the second advance point. If so, control A to move to the evasion position in each first position, or in other words, control A to walk in each evasion position, and when B moves to any second position, control A to move to the first advance point and control B to move to the second advance point (that is, A and B enter the advance point together after they are all in place, at this time at most one friendly team member is attacked by X), and after they reach their respective advance points, they are controlled to attack the enemy target together.

[0096] On this basis, optionally, taking A as an example, the movement path of A can be determined according to the movement coordinates of A and based on the A* algorithm. Further optionally, in the case that the movement coordinates cannot be determined after each step (that is, the attack target, the pursuit target, or the treasure to be found cannot be determined), if it is still outside the safe area, it can enter the safe area based on the shortest path; if it is already in the safe area, it can go to the center point of the safe area; if the safe area no longer exists, it can go to the original point coordinates; and the like, which are not specifically limited here.

[0097] Embodiment Four

[0098] Figure 5 The structure block diagram of the object action control device provided for the fourth embodiment of the present application is provided, and the device is used to execute the object action control method provided by any of the above embodiments. The device and the object action control method of each embodiment belong to the same inventive concept, and the details not described in detail in the embodiment of the object action control device can be referred to the embodiment of the object action control method. Referring to Figure 5 , the device can specifically include: a model acquisition module 410, a first action prediction module 420, a second action prediction module 430, and an action control module 440.

[0099] The model obtaining module 410 is configured to obtain a first action prediction model of a first object, an action object, and a second action prediction model of a second object belonging to a same object team as the first object.

[0100] The first action prediction module 420 is configured to predict, based on the first action prediction model, whether the second object has an action demand for the action object, and if so, predict whether the action can be successfully performed when the first object and the second object jointly perform the action on the action object.

[0101] The second action prediction module 430 is configured to, if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully perform the action on the action object, predict, based on the second action prediction model, whether the second object has an action demand for the action object, and if so, predict whether the action can be successfully performed when the first object and the second object jointly perform the action on the action object.

[0102] The action control module 440 is configured to, when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform the action on the action object, control the object combination to perform the action on the action object.

[0103] Optionally, the object action control device can further include:

[0104] The first action demand prediction module is configured to, if it is determined that the action cannot be successfully performed based on the first action prediction model, obtain a third object belonging to a same object team as the first object, and predict whether the third object has an action demand for the action object.

[0105] The first action success prediction module is configured to, if so, predict whether the action can be successfully performed when the first object, the second object, and the third object jointly perform the action on the action object.

[0106] The second action prediction module 430 can include:

[0107] The second action success prediction unit is configured to, if it is determined according to the prediction result of the first action prediction model that the object combination including the first object, the second object, and the third object can successfully perform the action on the action object.

[0108] The object action control device can further include:

[0109] The second action demand prediction module is configured to, if it is determined that the action cannot be successfully performed based on the second action prediction model, predict whether the third object has an action demand for the action object.

[0110] The second action success prediction module is configured to, if so, predict whether the action can be successfully performed when the first object, the second object, and the third object jointly perform the action on the action object.

[0111] The action control module 440 is specifically configured to:

[0112] When it is determined according to the prediction result of the second action prediction model that the object combination can successfully act on the action object, the third action prediction model of the third object is used to predict whether the third object has an action demand for the action object, and if so, it is predicted whether the action can be successfully performed when the first object, the second object and the third object jointly act on the action object.

[0113] When it is determined according to the prediction result of the third action prediction model that the object combination can successfully act on the action object, the object combination is controlled to act on the action object.

[0114] Optionally, the object action control device can further include:

[0115] The object association module is configured to, after predicting whether the action can be successfully performed based on the first action prediction model, if so, associate the action object with the first object.

[0116] The second action prediction module 430 can include:

[0117] The first action success prediction unit is configured to, if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully act on the action object, and the association result between the action object and the first object is obtained.

[0118] Optionally, the first action prediction module 420 and / or the second action prediction module 430 can include:

[0119] The second leading point prediction unit is configured to predict, from each leading point corresponding to the action object, a second leading point in which the second object is located when acting on the action object, wherein the leading point includes a position adjacent to the action position of the action object.

[0120] The action success prediction unit is configured to predict whether the action can be successfully performed when the first object is at the first leading point and the second object is at the second leading point, wherein the first leading point includes a leading point in which the first object is located when acting on the action object.

[0121] On this basis, optionally, the moving distance between the first object and the first leading point is less than the moving distance between the second object and the second leading point, a position adjacent to the first leading point and not adjacent to the action position is taken as the first position, and a position adjacent to the second leading point and not adjacent to the action position is taken as the second position.

[0122] The action control module 440 can include:

[0123] The first mobile control unit is configured to control the first object to move towards the first leading point and control the second object to move towards the second leading point.

[0124] The avoidance determination unit is configured to determine whether the first object needs to avoid the action object when the first object moves to any first position.

[0125] The second mobile control unit is configured to determine an avoidance position from the first positions if so, and control the first object to move towards the avoidance position.

[0126] The third mobile control unit is configured to control the first object to move to the first leading point and control the second object to move to the second leading point when the second object moves to any second position.

[0127] The action control unit is configured to control the object combination including the first object and the second object to act on the action object when the first object reaches the first leading point and the second object reaches the second leading point.

[0128] Optionally, the action object includes an attack object, and the object action control device further includes:

[0129] The to-be-acquired item determination module is configured to acquire candidate to-be-acquired items within a preset visual field range of the first object if it is determined according to the prediction result of the first action prediction model that the object combination cannot attack the attack object successfully, and determine whether there is a to-be-acquired item that needs to be acquired from the candidate to-be-acquired items according to a need degree and an acquisition distance of the first object to each candidate to-be-acquired item.

[0130] The object action determination module is configured to acquire a pursuit object of the first object if so, and determine whether to control the first object to pursue the pursuit object or acquire the to-be-acquired item according to a pursuit distance between the first object and the pursuit object, an acquisition distance between the first object and the to-be-acquired item, and a priority coefficient of the to-be-acquired item relative to the pursuit object.

[0131] Optionally, the object action control device further includes:

[0132] The object existence determination module is configured to determine whether there is an object that can be acted on successfully by the first object alone from each candidate object within a preset range of the first object, wherein the preset range includes a preset action range or a preset visual field range.

[0133] The action object determination module is configured to determine the action object from each candidate object if not.

[0134] The object action control device provided in the fourth embodiment of the present application comprises a model acquisition module, a first action prediction module, a second action prediction module and an action control module. The model acquisition module is configured to acquire a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to a same object team as the first object. The first action prediction module is configured to predict, based on the first action prediction model, whether the second object has an action demand for the action object, and if so, predict whether the action can be successfully performed when the first object and the second object jointly perform the action on the action object. The second action prediction module is configured to, if it is determined according to the prediction result of the first action prediction model that the object combination comprising the first object and the second object can successfully perform the action on the action object, i.e., when it is determined that the first object performs the action on the action object, predict again, based on the second action prediction model, whether the second object has an action demand for the action object, and if so, predict whether the action can be successfully performed when the first object and the second object jointly perform the action on the action object. The action control module is configured to control the object combination to perform the action on the action object when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform the action on the action object. The above device predicts the action of the object itself and / or the action of the object not determined to have the action, which belongs to the same object team as the object itself, by the action prediction model of the object, and controls each object in the object combination to perform the action on the action object in team cooperation when the object combination comprising the objects corresponding to the at least two action prediction models is determined to successfully perform the action on the action object by the at least two action prediction models, thereby achieving the effect of cooperative action between objects.

[0135] The object action control device provided in the embodiments of the present application can execute the object action control method provided in any of the embodiments of the present application, and has the function modules and beneficial effects corresponding to the execution method.

[0136] It should be noted that in the embodiments of the above object action control device, each unit and module included is only divided according to the function logic, but is not limited to the above division, as long as the corresponding function can be realized; in addition, the specific names of each functional unit are only for the convenience of mutual differentiation, and are not used to limit the protection scope of the present application.

[0137] Embodiment five

[0138] Figure 6 The structural schematic diagram of the object action control device provided in the fifth embodiment of the present application is shown in Figure 6 The device comprises a memory 510, a processor 520, an input device 530 and an output device 540. The number of processors 520 in the device can be one or more, Figure 6 and the memory 510, the processor 520, the input device 530 and the output device 540 in the device can be connected through a bus or other means, Figure 6The bus 550 is connected by way of example.

[0139] The memory 510, as a computer readable storage medium, can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the object action control method in the embodiments of the present application (for example, the model acquisition module 410, the first action prediction module 420, the second action prediction module 430 and the action control module 440 in the object action control device). The processor 520 executes various function applications and data processing of the device by running the software programs, instructions and modules stored in the memory 510, that is, implements the object action control method described above.

[0140] The memory 510 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and application programs required by at least one function; the data storage area can store data created according to the use of the device, etc. In addition, the memory 510 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state memory device. In some examples, the memory 510 can further include a memory remotely arranged with respect to the processor 520, which can be connected to the device through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network and a combination thereof.

[0141] The input device 530 can be used to receive input digital or character information, and generate key signal input related to user settings and function control of the device. The output device 540 can include a display device such as a display screen.

[0142] Embodiment six

[0143] Embodiment six of the present application provides a storage medium containing computer executable instructions, which when executed by a computer processor, are used to perform an object action control method, the method comprising:

[0144] Obtaining a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to the same object team as the first object;

[0145] Based on the first action prediction model, it is predicted whether the second object has an action demand for the action object, and if so, it is predicted whether the action can be successfully performed when the first object and the second object jointly perform the action on the action object;

[0146] If it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully act on the action object, whether the second object has an action demand on the action object is predicted based on the second action prediction model, and if yes, whether the action can be successfully performed when the first object and the second object jointly act on the action object is predicted.

[0147] When it is determined according to the prediction result of the second action prediction model that the object combination can successfully act on the action object, the object combination is controlled to act on the action object.

[0148] Of course, the storage medium provided by the embodiment of the present application includes computer executable instructions, which are not limited to the method operations described above, and can also perform related operations in the control method of object action provided by any embodiment of the present application.

[0149] Through the above description of the embodiments, those skilled in the art can clearly understand that the present application can be realized by means of software and necessary general hardware, and of course can also be realized by hardware, but in many cases the former is a better embodiment. According to such understanding, the technical solutions of the present application or the part that contributes to the prior art can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a floppy disk, a read-only memory (ROM), a random access memory (RAM), a FLASH, a hard disk or an optical disk, etc., including a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the method described in various embodiments of the present application.

[0150] Note that the above is only the preferred embodiment of the present application and the technical principle applied. Those skilled in the art will understand that the present application is not limited to the specific embodiments described herein, and those skilled in the art can make various obvious changes, readjustments and substitutions without departing from the scope of the present application. Therefore, although the present application has been described in more detail through the above embodiments, the present application is not limited to the above embodiments, and can include more other equivalent embodiments without departing from the concept of the present application, and the scope of the present application is determined by the scope of the appended claims.

Claims

1. A control method of an object action, characterized by, The method comprises: obtaining a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to a same object team as the first object; predicting whether the second object has an action demand for the action object based on the first action prediction model, and if so, predicting whether an action can be successfully performed when the first object and the second object jointly perform an action on the action object; if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully perform an action on the action object, predicting whether the second object has an action demand for the action object based on the second action prediction model, and if so, predicting whether an action can be successfully performed when the first object and the second object jointly perform an action on the action object; controlling the object combination to perform an action on the action object when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform an action on the action object.

2. The method of claim 1, wherein, After predicting whether an action can be successfully performed based on the first action prediction model, the method further comprises: if not, obtaining a third object belonging to the same object team as the first object, and predicting whether the third object has an action demand for the action object; if so, predicting whether an action can be successfully performed when the first object, the second object, and the third object jointly perform an action on the action object; if it is determined according to the prediction result of the first action prediction model that the object combination including the first object, the second object, and the third object can successfully perform an action on the action object; after predicting whether an action can be successfully performed based on the second action prediction model, the method further comprises: if not, predicting whether the third object has an action demand for the action object; if so, predicting whether an action can be successfully performed when the first object, the second object, and the third object jointly perform an action on the action object; controlling the object combination to perform an action on the action object when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform an action on the action object. after predicting whether an action can be successfully performed based on the first action prediction model, the method further comprises: if it is determined according to the prediction result of the first action prediction model that the object combination including the first object and the second object can successfully perform an action on the action object, predicting whether the second object has an action demand for the action object based on the second action prediction model, and if so, predicting whether an action can be successfully performed when the first object and the second object jointly perform an action on the action object; controlling the object combination to perform an action on the action object when it is determined according to the prediction result of the second action prediction model that the object combination can successfully perform an action on the action object.

3. The method of claim 1, wherein, after predicting whether an action can be successfully performed based on the first action prediction model, the method further comprises: If yes, the action object is associated with the first object; The determining whether the object combination including the first object and the second object can act on the action object successfully according to the prediction result of the first action prediction model includes: If the object combination including the first object and the second object can act on the action object successfully according to the prediction result of the first action prediction model, and the association result between the action object and the first object is obtained.

4. The method of claim 1, wherein, The prediction includes whether the first object and the second object can act on the action object successfully. The second object is predicted to be at a second preceding point when the second object acts on the action object from each preceding point corresponding to the action object, wherein the preceding point includes a position adjacent to an action position of the action object. The prediction includes whether the first object and the second object can act on the action object successfully when the first object is at a first preceding point and the second object is at a second preceding point.

5. The method of claim 4, wherein, The moving distance between the first object and the first preceding point is less than the moving distance between the second object and the second preceding point, and a first position adjacent to the first preceding point and not adjacent to the action position is set as the first position, and a second position adjacent to the second preceding point and not adjacent to the action position is set as the second position. The controlling the object combination to act on the action object includes: The first object is controlled to move towards the first preceding point, and the second object is controlled to move towards the second preceding point. When the first object moves to any of the first positions, it is determined whether the first object needs to be controlled to avoid the action object. If yes, an avoidance position is determined from each of the first positions, and the first object is controlled to move towards the avoidance position. When the second object moves to any of the second positions, the first object is controlled to move to the first preceding point, and the second object is controlled to move to the second preceding point. When the first object reaches the first preceding point and the second object reaches the second preceding point, the object combination including the first object and the second object is controlled to act on the action object.

6. The method of claim 1, wherein, The action object includes an attack object, and the method further includes: If the object combination cannot attack the attack object successfully according to the prediction result of the first action prediction model, candidate acquisition objects within a preset visual field range of the first object are obtained, and whether there is a to-be-acquired object that needs to be acquired in each of the candidate acquisition objects is determined according to a need degree and an acquisition distance of the first object to each of the candidate acquisition objects. If yes, a pursuit object of the first object is acquired, and a pursuit distance between the first object and the pursuit object, the acquisition distance between the first object and the to-be-acquired object, and a priority coefficient of the to-be-acquired object relative to the pursuit object are used to determine whether the first object pursues the pursuit object or acquires the to-be-acquired object.

7. The method of claim 1, wherein, Further comprising: For each candidate object in a preset range of the first object, it is determined whether there is an object in each candidate object that can be successfully acted on by the first object, wherein the preset range includes a preset action range or a preset field of view range. If no, the action object is determined from each candidate object.

8. An object action control device characterized by comprising: Comprising: a model acquisition module, configured to acquire a first action prediction model of a first object and an action object, and a second action prediction model of a second object belonging to a same object team as the first object; a first action prediction module, configured to predict, based on the first action prediction model, whether the second object has an action demand for the action object, and if yes, predict whether an action can be successfully performed when the first object and the second object jointly perform an action on the action object; a second action prediction module, configured to predict, based on the second action prediction model, whether the second object has an action demand for the action object if it is determined according to a prediction result of the first action prediction model that an object combination including the first object and the second object can successfully perform an action on the action object, and if yes, predict whether an action can be successfully performed when the first object and the second object jointly perform an action on the action object; an action control module, configured to control the object combination to perform an action on the action object if it is determined according to a prediction result of the second action prediction model that the object combination can successfully perform an action on the action object.

9. An object action control device characterized by comprising: Comprising: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the object action control method according to any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the object action control method according to any one of claims 1-7.

Images