18194116. PHYSICS-BASED SIMULATION OF HUMAN CHARACTERS IN MOTION simplified abstract (NVIDIA Corporation)

PHYSICS-BASED SIMULATION OF HUMAN CHARACTERS IN MOTION

Organization Name

NVIDIA Corporation

Inventor(s)

Zhengyi Luo of Pittsburgh PA (US)

Jason Peng of Vancouver (CA)

Sanja Fidler of Toronto (CA)

Or Litany of Sunnyvale CA (US)

Davis Winston Rempe of Redwood City CA (US)

Ye Yuan of Santa Clara CA (US)

PHYSICS-BASED SIMULATION OF HUMAN CHARACTERS IN MOTION - A simplified explanation of the abstract

This abstract first appeared for US patent application 18194116 titled 'PHYSICS-BASED SIMULATION OF HUMAN CHARACTERS IN MOTION

Simplified Explanation

The patent application describes a system for generating a simulated environment and updating a machine learning model to control the movement of human characters with different body shapes within the environment based on their respective characteristics.

• The system involves generating a simulated environment where human characters with diverse characteristics can interact and move according to their body shapes.
• A machine learning model is used to determine actions for each human character based on their humanoid state, body shape, and task-related features such as environmental factors and trajectories.

Potential Applications

This technology could be applied in the development of virtual reality games, crowd simulation software for urban planning, and training simulations for emergency response scenarios.

Problems Solved

This technology solves the problem of creating realistic and dynamic simulations of human movement in various environments based on individual body shapes and characteristics.

Benefits

The system allows for more accurate and lifelike simulations of human behavior in virtual environments, leading to improved training scenarios and more realistic crowd simulations.

Potential Commercial Applications

Potential commercial applications include the development of virtual reality games, training simulations for various industries, and software for urban planning and crowd management.

Possible Prior Art

One possible prior art could be existing crowd simulation software that focuses on generic human movement patterns rather than individualized behaviors based on body shapes and characteristics.

How does the machine learning model adapt to new environments or scenarios?

The patent application does not provide details on how the machine learning model can be updated or retrained to adapt to new environments or scenarios. This could be a crucial aspect of the technology's practical implementation.

What kind of data is required to train the machine learning model for accurate human character movement?

The patent application does not specify the type or amount of data needed to train the machine learning model effectively for controlling human character movement in simulated environments. Understanding the data requirements could be essential for implementing this technology successfully.

Original Abstract Submitted

In various examples, systems and methods are disclosed relating to generating a simulated environment and update a machine learning model to move each of a plurality of human characters having a plurality of body shapes, to follow a corresponding trajectory within the simulated environment as conditioned on a respective body shape. The simulated human characters can have diverse characteristics (such as gender, body proportions, body shape, and so on) as observed in real-life crowds. A machine learning model can determine an action for a human character in a simulated environment, based at least on a humanoid state, a body shape, and task-related features. The task-related features can include an environmental feature and a trajectory.