Font Size: a A A

Based On Virtual Augmented Reality Interactive Technology Research And Application Of The Proposed Avatar

Posted on:2020-06-23Degree:MasterType:Thesis
Country:ChinaCandidate:Z N ZhangFull Text:PDF
GTID:2438330623964238Subject:Computer technology
Abstract/Summary:PDF Full Text Request
With the progress of augmented reality technology,interactive technology in augmented reality scene has gradually become one of the development directions of next generation human-computer interaction.In the multi-person interaction system of augmented reality,in order to realize the interaction between users and virtual interactive objects,the system establishes virtual avatars for each user separately.As the agent of users in the system,virtual avatars can replace users to interact with virtual interactive objects and complete the interactive tasks.Based on virtual avatar motion modeling technology,the interactive semantic recognition and interactive intention understanding technology in combat training scenarios are studied in this paper.In order to establish a hierarchical virtual avatar model,we analyze the structure and movement characteristics of human bones.Through virtual avatar posture calibration algorithm and bone length estimation algorithm,the virtual body initialization calibration is completed,we establish a precise virtual avatar model.Virtual avatar motion modeling technology is mainly realized by human attitude sensing based on inertial sensor and human body location based on optical sensor.On this basis,a virtual avatar motion retarget algorithm based on multi-spatial position constraints is adopted to avoid the distortion of virtual avatar motion.Aiming at the combat training interaction in augmented reality scenarios,the definition of interaction semantics and interaction intent is elaborated,and the interactive semantics mapping of eight human interaction actions is carried out.In order to identify the interactions of the trainers,the human interaction data is collected and preprocessed,and the three-layer Bi LSTM model is used to learn and identify them.According to the trainer interaction sequence data,we used the classic Bayesian network based on prior knowledge to understand the interaction intention of the trainers.On this basis,we designed and constructed the combat training interaction scenario,and verified examples.
Keywords/Search Tags:Virtual Avatar, Motion Modeling, Interactive Recognition, Intention Understanding
PDF Full Text Request
Related items