Generating motion markers for simulating traffic using machine learning models
By adopting a transformer-based artificial intelligence model in traffic modeling, marking the movement of the agent and training/updating the model, the problems of poor realism and scalability in traditional methods are solved, and efficient traffic simulation is achieved.
CN119723487APending Publication Date: 2025-03-28NVIDIA CORP
0 Cites 0 Cited by
Patent Information
- Application Number
- CN202411362652.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2024-05-15
- Filing Date
- 2024-09-27
- Publication Date
- 2025-03-28
AI Technical Summary
Technical Problem
Traditional traffic modeling methods are difficult to achieve realism and scalability in complex environments, resulting in poor simulation performance.
Method used
Transportation modeling is performed using converter-based artificial intelligence models, and the transformer model is trained/updated to simulate traffic movements of different agents over time.
Benefits of technology
It realizes scalable and realistic traffic simulation of multiple agents in complex environments, improving simulation performance.
✦ Generated by Eureka AI based on patent content.
Abstract
The invention discloses generating motion markers for simulating traffic using a machine learning model. In various examples, systems and methods related to generating markers for traffic modeling are specifically disclosed. One or more circuits may identify a trajectory in the dataset and generate an action according to the identified trajectory. The one or more circuits may generate, based at least on the plurality of actions and at least one trajectory of the plurality of trajectories, a set of markers representing the actions to generate trajectories of the one or more agents in the simulation. The one or more circuits may update the converter model to generate a simulated action of the simulated agent based at least on a marker generated from a trajectory in the dataset.
Need to check novelty before this filing date? Find Prior Art