Emotional Communication Robot Based on 3D Face Model and ASR Technology

Today, some robots have the ability to express emotions, which makes human-robot interaction (HRI) more realistic. However, most of the current robots do not have real face, and people's facial expressions are very important in communication. Therefore, this paper constructs an emotional communication humanoid robot system based on 3D face and automatic speech recognition (ASR) system. Chinese speech recognition is performed by an ASR system. Audio data can be used for other research. In order to express facial expressions and pronunciations more accurately and realistically, the real-life data collected by the OptiTrack system is used as a support, and weighted Dirichlet free-form deformations (DFFD) is applied to deform the 3D face model. ASR selects HMM-GMM model as the acoustic model and N-gram model as the language model. Acoustic features are selected as perceptual linear prediction (PLP) features.