隐式情绪导向的语音驱动仿生机器人说话方法

打开文本图片集

关键词：仿生机器人；音频驱动；情绪导向；卷积长短时记忆网络

中图分类号：TP391 文献标志码：A 文章编号：1001-3695（2025）08-007-2297-07

doi：10.19734/j.issn.1001-3695.2025.01.0011

Implicit emotion-oriented approach to speech-driven bionic robot facial expressions

XuKang，Yuan Ye，Fu Junxiu，FuKeting，RenQinze，Liu Na† （InstituteofMachineInteligence，UniversityofShanghaiforScience&Technology，Shanghai2Ooo93，China）

Abstract：Thisstudyproposedaninovativeimplicitemotion-orientedspeech-drivenmethodforgeneratingfacialexpressions andhead movements inbionicrobots.Traditional methodsrelyonpre-programmedrandomactionsequences，whichare insufficient for precise emotional expresion.Basedondeep leaing，thisstudyintroducedneck servocontrolcoeffcients inadition to facial expresions，achieving precise mappng fromaudiosignals tonatural expressions.Furthermore，thisstudydesignedanimplicitemotion-oriented featurefusionautoencoderframework thatcouldinferemotionalfeaturesimplicitlyfromaudio withoutrequiring explicit emotionparameters.Itenabledthegenerationof richfacialexpresions andneck movements that matchedtheaudiocontentandemotionalnuances.Experiments demonstrate thatthe proposed methodsignificantlyoutperforms existingtechnologies onmultiple datasets.Aditionall，through lightweight design，the proposedmethodeficientlyadapts to mobile devices with limited resources.

Key Words：biomimeticrobots；speech-driven；emotion-oriented;convolutional long short-term memory network

0 引言

随着人工智能技术的不断进步，仿生机器人在模拟人类行为和表情方面的能力日益增强，为多个领域带来了革命性的应用前景，尤其是在主动健康、社交互动、教育娱乐等方面。（剩余17402字）