AI Movement Control via Recorded Trajectories
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for controlling the movement of AI models in racing games require significant computing and processing overhead, leading to high power consumption on terminals.
Innovation Solution
The method involves obtaining recorded trajectory information from a real user's gameplay and using this information to control the movement of AI models, reducing the need for complex logic decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If logic decision-making is used to control AI model movement in real-time, then the AI model can make dynamic racing decisions, but the terminal requires large amounts of computing and processing overhead resulting in serious power consumption
Solution Approach 1:
The patent applies preliminary action by pre-recording trajectory information from human players and pre-processing it into control instructions. Instead of calculating AI movements in real-time during gameplay, the system prepares and stores optimal trajectories beforehand, then simply plays back the pre-computed path during the actual race, eliminating the need for continuous complex calculations on the terminal device.
Solution Approach 2:
The patent uses copying by replicating human player trajectories to control AI models. Rather than creating entirely new AI decision-making paths through complex algorithms, the system copies and adapts proven effective human racing trajectories, reducing computational requirements while maintaining realistic and effective AI behavior.
2Adaptability or versatility
If real-time logic decision-making is used to control AI model position, speed, and orientation, then the AI can adapt to changing race conditions, but the terminal experiences serious power consumption
Solution Approach 1:
The system performs adaptability analysis and trajectory generation in advance during an offline phase, where human player data is collected and processed to create comprehensive trajectory libraries. During actual gameplay, the system simply retrieves and plays back pre-computed trajectories that were already adapted to various race conditions, eliminating real-time computational energy consumption.
3Ease of operation
If complex logic decision-making algorithms are implemented for AI control, then the AI model can exhibit sophisticated racing behavior, but the terminal requires large computing resources
Solution Approach 1:
The patent simplifies complex AI control algorithms by copying and adapting human player control patterns. Instead of implementing sophisticated autonomous decision-making algorithms, the system records and replays human player inputs and trajectory decisions, achieving precise and natural AI behavior with significantly reduced computational complexity.
Solution Approach 2:
The patent extracts the complex computational tasks from the terminal device and relocates them to an external server or cloud platform during the offline trajectory generation phase. The terminal device only needs to store and play back the simplified trajectory instructions, separating the heavy computing workload from the actual gameplay execution.
Data Source
AI summary
A method for controlling movement of a virtual object in a virtual environment is performed at a terminal. The method includes: displaying a racing battle interface, the racing battle interface presenting a second virtual object and a third virtual object in a virtual scene of the virtual environment, the second virtual object being controlled by a user account associated with a user of the terminal; controlling movement of the third virtual object according to first recorded trajectory information in the racing battle, wherein the first recorded trajectory information represents a movement trajectory of a first virtual object in the virtual scene of the virtual environment; in accordance with a determination that the third virtual object has an abnormal condition, controlling the third virtual object to enter a free performance state according to the abnormal condition; and resuming the control movement of the third virtual object according to the first recorded trajectory information after the free performance state ends.