Search

CN-122006253-A - Data-driven game role action automatic optimization method and system

CN122006253ACN 122006253 ACN122006253 ACN 122006253ACN-122006253-A

Abstract

The invention provides a data-driven game character action automatic optimization method and a data-driven game character action automatic optimization system, which relate to the technical field of data processing, wherein the method comprises the steps of collecting scene data of a reference action space body according to real-time running diagram coordinates of a game character to obtain environment geometric parameters and environment physical attributes; constructing environment constraint vectors, receiving user real-time operation streams to perform standard action stream matching, executing constraint adaptive deformation, outputting optimized action streams, driving action rendering of game characters by adopting the optimized action streams, and outputting visual character scene actions. The invention solves the technical problems that the game role actions in the prior art are usually controlled through a predefined animation library or animation sequence and cannot be adjusted in real time according to environmental changes, thereby influencing the fluency of the game and the immersion of a player.

Inventors

  • LIU XIAOYUE

Assignees

  • 深圳超月文化传媒有限公司

Dates

Publication Date
20260512
Application Date
20251210

Claims (10)

  1. 1. A method for automatically optimizing the actions of a data-driven game character, the method comprising: According to the real-time running diagram coordinates of the game characters, scene data acquisition of the reference action space body is carried out, and the environment geometric parameters and the environment physical properties are obtained; Constructing an environment constraint vector according to the environment geometric parameters and the environment physical attributes; Receiving a user real-time operation flow to perform standard action flow matching; Executing constraint adaptive deformation on the standard action flow according to the environment constraint vector, and outputting an optimized action flow; And driving action rendering of the game characters by adopting the optimized action flow, and outputting visual character scene actions.
  2. 2. The data driven game character action automatic optimization method of claim 1, further comprising: Performing environment scene aggregation on the game map to obtain a typical game environment; performing action capturing and tracking in the process of driving the game characters to execute a full action library by the typical game environment to obtain full scene action capturing data; constructing an dynamic capture coordinate system by taking a role pelvis as an origin; And overlapping the full scene dynamic capture data in the dynamic capture coordinate system space, and constructing the reference action space body.
  3. 3. The data-driven game character motion automatic optimization method of claim 2, wherein the overlapping of the full scene dynamic capture data in the dynamic capture coordinate system space, constructing the reference motion space volume, comprises: Aggregating the full scene dynamic capture data according to the action types to obtain a plurality of dynamic capture data sets of a plurality of action types; Predefining multi-scale convex hull granularity according to the action amplitude attribute of the multiple action types; After the plurality of dynamic capture data sets are aligned and overlapped to the dynamic capture coordinate system, performing sub-action convex hull calculation according to the multi-scale convex hull granularity to obtain a plurality of hierarchical convex hull grids; And superposing and fusing the plurality of hierarchical convex hull grids through voxels to obtain the reference action space body.
  4. 4. The data driven game character action automatic optimization method of claim 1, wherein receiving a user real-time operation flow for standard action flow matching comprises: aligning the user real-time operation flow with a game engine clock to obtain an operation instruction sequence; extracting an instruction time stamp sequence from the operation instruction sequence; Traversing the instruction time stamp sequence to carry out false touch signal extraction to obtain a continuous input sequence; Encoding the continuous input sequence to obtain an operation feature vector; and after the operation feature vector is adopted to match and call a standard action template sequence in the pre-classification action template library, the standard action template sequence is smoothly spliced to generate the standard action flow.
  5. 5. The data driven game character action automatic optimization method of claim 4, further comprising: interactively obtaining a theoretical input time sequence and a theoretical hardware operation parameter sequence of the continuous input sequence; Comparing the theoretical input time sequence with the real operation input time sequence of the continuous input sequence to calculate an operation time error; Comparing the theoretical hardware operation parameter sequence with the real operation hardware operation parameter sequence of the continuous input sequence to calculate an operation space error; Performing operation tolerance correction on the continuous input sequence according to the operation time error and the operation space error to obtain a corrected input sequence; and performing time sequence smooth splicing of the standard action template sequence according to the corrected input sequence to generate the standard action flow.
  6. 6. The data driven game character action automatic optimization method of claim 1, wherein constructing an environment constraint vector from the environment geometric parameters and environment physical properties comprises: Traversing the environment geometric parameters to extract the distance of the obstacle by taking the real-time running map coordinates as an action origin, and generating a radial distance vector; Extracting a vertical distance vector from the environmental geometric parameter according to the real-time running map coordinates; decomposing the environmental physical attribute in a horizontal level to obtain a local ground physical vector and a local environmental physical vector; and packaging the radial distance vector, the vertical distance vector, the local ground physical property vector and the local environment physical property vector to obtain the environment constraint vector.
  7. 7. The data-driven game character action automatic optimization method of claim 6, wherein performing constraint-adaptive deformation on the standard action flow according to the environment constraint vector, outputting an optimized action flow, comprising: performing track previewing on the standard action flow, and positioning a plurality of key space-time nodes; adopting the environment constraint vector to execute constraint adaptive deformation correction on the plurality of key space-time nodes to obtain a plurality of deformation optimization action fragments; And inserting the deformation optimization action fragments into the standard action flow in a covering manner, and outputting the optimization action flow.
  8. 8. The data driven game character motion automatic optimization method of claim 7, wherein performing constraint-adaptive deformation correction at the plurality of key spatiotemporal nodes using the environmental constraint vector to obtain a plurality of deformation-optimized motion segments, comprising: performing action horizontal amplitude convergence analysis based on the radial distance vector to generate a horizontal convergence coefficient; Performing vertical gesture adaptation analysis based on the vertical distance vector to generate a bowing factor; Performing action inertia correction analysis based on the local ground physical property vector, and outputting an inertia correction amount; performing role confidence offset analysis according to the local environment physical property vector to generate equivalent offset force; And performing constraint adaptive deformation joint correction on the key space-time nodes by adopting the horizontal convergence coefficient, the bowing factor, the inertia correction amount and the equivalent offset force to obtain a plurality of deformation optimization action fragments.
  9. 9. The data-driven game character action automatic optimization method of claim 1, wherein scene physical interaction feedback rendering is synchronously performed during action rendering of the game character driven by the optimized action stream, and visual scene interaction actions are output.
  10. 10. A data-driven game character action automatic optimization system for implementing the data-driven game character action automatic optimization method according to any one of claims 1 to 9, the system comprising: The data acquisition module is used for acquiring scene data of the reference action space body according to the real-time running diagram coordinates of the game character to obtain environment geometric parameters and environment physical properties; The constraint vector construction module is used for constructing an environment constraint vector according to the environment geometric parameters and the environment physical attributes; The action flow matching module is used for receiving the real-time operation flow of the user to perform standard action flow matching; The adaptive deformation module is used for executing constraint adaptive deformation on the standard action flow according to the environment constraint vector and outputting an optimized action flow; And the action rendering module is used for driving action rendering of the game characters by adopting the optimized action flow and outputting visual character scene actions.

Description

Data-driven game role action automatic optimization method and system Technical Field The invention relates to the technical field of data processing, in particular to a data-driven game role action automatic optimization method and system. Background With the rapid development of modern game technology, the action performance of game characters has become one of the key factors affecting the immersion of games and player experience, and in order to make the behavior of game characters more natural, real and interactive, game developers often need to develop complex animation systems and physical engines to drive character actions. In conventional games, the actions of characters are usually controlled by predefined animation libraries or animation sequences, which are static, the actions and reactions of the characters are preset, cannot be adjusted in real time according to environmental changes, and if the characters enter complex or irregular environments, such as uneven ground, narrow space, etc., the actions of the characters become unnatural or cannot be performed, for example, the characters may catch, jump to an inappropriate height or collide with obstacles, which seriously affects the smoothness of the game and the immersion of the players. Disclosure of Invention The application provides a data-driven game role action automatic optimization method and a data-driven game role action automatic optimization system, and aims to solve the technical problems that in the prior art, game role actions are usually controlled through a predefined animation library or an animation sequence and cannot be adjusted in real time according to environmental changes, so that the smoothness of a game and the immersion of a player are affected. The application discloses a first aspect, which provides a data-driven game role action automatic optimization method, which comprises the steps of collecting scene data of a reference action space body according to real-time running diagram coordinates of a game character to obtain environment geometric parameters and environment physical attributes, constructing environment constraint vectors according to the environment geometric parameters and the environment physical attributes, receiving a user real-time operation flow to perform standard action flow matching, executing constraint adaptive deformation on the standard action flow according to the environment constraint vectors to output an optimized action flow, and driving action rendering of the game character by adopting the optimized action flow to output visual role scene actions. The application discloses a second aspect, which provides a data-driven game role action automatic optimization system, which is used for the data-driven game role action automatic optimization method, and comprises a data acquisition module, a constraint vector construction module, an action flow matching module, an adaptation type deformation module and an action rendering module, wherein the data acquisition module is used for acquiring scene data of a reference action space body according to real-time running diagram coordinates of a game figure to obtain environment geometric parameters and environment physical properties, the constraint vector construction module is used for constructing environment constraint vectors according to the environment geometric parameters and the environment physical properties, the action flow matching module is used for receiving a user real-time operation flow to perform standard action flow matching, the adaptation type deformation module is used for executing constraint adaptation type deformation on the standard action flow according to the environment constraint vectors to output optimized action flow, and the action rendering module is used for driving action rendering of the game figure to output visual role scene action. The one or more technical schemes provided by the application have at least the following beneficial effects: The method comprises the steps of acquiring scene data according to real-time running diagram coordinates of game characters, acquiring environment geometric parameters and environment physical attributes, enabling behaviors of game characters to reflect states of the environments where the game characters are located in real time, enabling optimization and adaptability of character actions to be stronger, constructing environment constraint vectors through analysis of the environment geometric parameters and the physical attributes, effectively converting environment factors into limiting conditions of character motions, accordingly guaranteeing that the character actions do not collide or violate the behaviors of physical laws in complex environments, receiving real-time operation flows of users and matching with standard action flows in game engines, converting user inputs into preset standard actions, enabling the input of players to be accurately mapped into the act