Human skeleton model-based personnel action real-time simulation method

Abstract

The invention provides a human skeleton model-based personnel action real-time simulation method. The method mainly comprises four steps of constructing an initial transformation matrix based on a human skeleton model, converting motion data, constructing a real-time transformation matrix based on personnel action data and performing personnel action real-time simulation. The construction of the initial transformation matrix based on the human skeleton model is divided into the construction of a three-dimensional human skeleton model and the construction of the initial transformation matrix; motion data conversion is divided into matrix-to-four element conversion and coordinate system conversion; the construction of a real-time transformation matrix based on the personnel action data is divided into absolute rotation amount calculation, absolute displacement amount calculation and real-time transformation matrix construction; and the action real-time simulation is based on the constructed initial transformation matrix and real-time transformation matrix, and the position and posture of the constructed human body model in the three-dimensional space are calculated in each frame by utilizing the calculation capability of a three-dimensional simulation engine. According to the method, the accuracy and real-time performance of three-dimensional scene synchronous simulation can be improved.

Description

technical field [0001] The invention belongs to the field of real-time simulation of personnel movements, in particular to a method for real-time simulation of personnel movements based on a human skeleton model. Background technique [0002] With the development of spatial positioning technology and wearable devices, as well as the continuous improvement of immersive application requirements, more and more simulation applications require a certain number of people to participate in the simulation scene in a "human-in-the-loop" manner. The virtual personnel are generated synchronously in the 3D virtual simulation environment, and the behavior characteristics of the virtual personnel are consistent with those of the participants. The "human-in-the-loop" method is that personnel participate in the simulation process, and at the same time a virtual personnel model is generated in the 3D virtual simulation environment. Participants can control the survival time and behavior of t...

AI Technical Summary

Problems solved by technology

[0004] The drawing of the 3D human body model is more complicated. The more bones of the human body model, the more motion data that needs to be processed when the 3D simulation engine presents the 3D human body model in the 3D virtual simulation environment when the participants are exercising. The larger the amount, the higher the amount of calculation will consume the performance of the computer, which will cause delays and freezes in the simulation scene, so that the actions of the participants cannot be simulated synchronously in the 3D virtual simulation environment in real time

[0006] (1) Motion capture technology is mostly used to make 3D animation, and its application fields are concentrated in the production of animation and games. The real-time simulation application of personnel combat based on motion capture technology started relatively late, and the existing motion capture based on 3D virtual engine real-time Simulation technology is still not up to the training requirements

Especially when multiple "human-in-the-loop" nodes are training at the same time, the 3D virtual engine will often be "stuck" due to excessive data calculation peaks, resulting in interruption of the training process

(2) When the existing virtual engine performs motion capture real-time simulation, generally only one "human-in-the-loop" node is connected. When multiple "human-in-the-loop" nodes work together, the motion simulation of the virtual engine is not accurate enough

(3) The simulation engine of the global coordinate system is used to obtain the offset and rotation of the moving bones in the motion capture data. The transformation matrix is ​​constructed in an iterative manner, and the relative offset and rotation in the motion data are converted into absolute ones, so that To control the movement of the bones, this method will perform a large number of matrix operations, occupying a lot of computer performance resources, which will lead to unsmooth motion simulation

Images