Research On Generation Of Expressive Teaching Gestures For Humanoid Robot

The teaching assistant robot is the product of using educational robots to assist classroom teaching.Teaching assistant robots with human-like voice and motion interaction capabilities play an important role in assisting the smooth development of teaching.In recent years,the development of artificial intelligence technology has greatly improved the voice interaction capabilities of robots.However,the research on the interactive actions that accompany speech,that is,the generation of co-speech gestures is still relatively scarce.The generated co-speech gestures still have problems such as mechanical rigidity and insufficient coherence.How to make the teaching assistant robot have the ability to autonomously generate human-like co-speech expressive teaching gestures and improve its body language communication level is a problem that needs to be solved urgently.Based on the above analysis,this article focuses on the generation of expressive teaching gestures of humanoid teaching assistant robots,aiming to improve the expressive teaching gesture generation ability of teaching assistant robots in terms of semantic relevance,gesture diversity and gesture continuity.The specific research content is as follows:(1)In the modeling of expressive teaching gestures,in order to construct the expressive teaching gesture model in the teaching scene,the expressive teaching gesture classification method and the expressive teaching gesture mining method were proposed.The related mining results formed the first Chinese expressiveness Teaching gesture data set.(2)On the generation of expressive teaching gestures based on instructional and descriptive short texts,a method for generating expressive teaching gestures based on keyword matching is proposed,which overcomes the semantic correlation between the gestures generated by traditional methods under short text conditions The problem of lack of sex.(3)In the context of semantically coherent long text expressive teaching gestures,a speech context-based expressive teaching gesture sequence generation method is proposed,which realizes that the robot can autonomously generate a coherent and semantically highly consistent expressive teaching gesture sequence.Finally,this paper designs and implements a humanoid robot expressive teaching gesture generation system,and uses Nao humanoid robot as a platform to verify the method in this paper.The experimental results show that the gestures generated by the method proposed in this paper not only ensure high relevance to text semantics,but also ensure the diversity of gestures,and also achieve better contextual coherence.The related results have certain reference significance for improving the body language expression ability of teaching assistant robots and promoting the social humancomputer interaction of teaching assistant robots.