Hierarchical Reinforcement Learning for Adaptive Virtual Object Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing virtual reality and augmented reality technologies are inefficient in rendering virtual objects that interact naturally with real environments in real time, failing to provide enhanced user experiences by reflecting changes in the real environment.
Innovation Solution
A processor-implemented method using hierarchical reinforcement learning to determine predictive trajectories and direction information of virtual objects based on Gaussian random paths, incorporating a high-level model for trajectory encoding and a low-level model for direction rendering, ensuring seamless interaction with real environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If typical rendering technology is used, then the system is simple to implement, but the virtual object cannot interact naturally with the real environment in real time
Solution Approach 1:
The system is divided into multiple modules including trajectory generation module, rendering module, and environment interaction module. Each module handles specific aspects of the virtual object rendering process, allowing complex natural interactions to be achieved through coordinated simple components rather than a monolithic complex system.
Solution Approach 2:
The system pre-generates multiple predictive trajectories for the virtual object before actual rendering occurs. By preparing these trajectories in advance based on environment predictions, the system enables real-time natural interaction without requiring complex real-time computation during the actual rendering phase.
2Adaptability or versatility
If real-time rendering of virtual objects is implemented, then user experience is enhanced, but computational efficiency decreases
Solution Approach 1:
The system generates multiple predictive trajectories beyond what is strictly necessary, selecting the most appropriate one based on actual environmental conditions. This approach allows efficient real-time rendering by preparing options in advance while only utilizing the necessary portion during actual execution, balancing adaptability with computational efficiency.
Solution Approach 2:
The patent replaces traditional physics-based mechanical simulation systems with machine learning models that predict trajectories and rendering parameters. This substitution eliminates computationally intensive real-time physics calculations while maintaining natural interaction capabilities through data-driven predictions.
3Adaptability or versatility
If virtual objects are rendered to reflect real environment changes, then user experience sharing is improved, but processing time increases
Solution Approach 1:
The system performs preliminary rendering of multiple predictive trajectories in advance, storing these pre-rendered options for quick selection and application. When environmental changes occur, the system simply selects and applies the appropriate pre-rendered trajectory rather than performing time-consuming real-time rendering calculations, thus maintaining accurate environment reflection with minimal processing time.
Solution Approach 2:
The system creates multiple copies of the virtual object along different predictive trajectories in advance. These copies are pre-rendered and stored, allowing the system to quickly switch between them based on actual environmental conditions without requiring time-consuming real-time rendering operations.
Data Source
AI summary
A processor-implemented method with virtual object rendering includes: determining a plurality of predictive trajectories of a first object according to a Gaussian random path based on a high-level model that is trained by hierarchical reinforcement learning; determining direction information of a second object according to subgoals corresponding to the predictive trajectories based on a low-level model that is trained by hierarchical reinforcement learning; determining direction information of the second object according to a subgoal corresponding to one of the predictive trajectories based on an actual trajectory of the first object; and rendering the second object, which is a virtual object, based on the determined direction information.