摘要
情感模型的建立,是产生人工情感的一种途径,也是和谐人机交互的基础.针对现有的隐马尔可夫情感模型仅能产生基本情绪的问题,提出了一种改进的模型,使其能够产生复合情绪.首先,放宽已有理论的假设条件,使得某种刺激可以引发多种基本情绪,并且2种基本情绪状态的强度可以同时增大,提高了模型的普适性;其次在引入辅助矩阵和可变阈值后,实现了情感模型的复合情绪生成.通过仿真试验,验证了该模型的有效性.
Emotion model building provides a way to generate artificial emotion.In addition,it is the foundation of natural human-robot interaction.In order to solve the problem of the existing hidden Markov model（HMM） only generating basic emotion,an improved model was proposed.It is capable of generating complex emotion.First,the hypothesis terms of the HMM emotion theory were softened,which made one stimuli trigger multiple emotions leading to a simultaneous increase in the intensity of two basic emotions,which universally enhanced the model.Second,an assistance matrix and variable threshold were introduced in order to implement complex emotion computing.Through experimentation,the effect of the model was verified.
出处
《智能系统学报》
2010年第6期545-550,共6页
CAAI Transactions on Intelligent Systems
基金
国家"863"计划资助项目(2007AA04Z218)
国家自然科学基金资助项目(60903067)
关键词
情感计算
人工心理
人工情感
隐马尔科夫情感模型
emotion computing
artificial psychology
artificial emotion
HMM emotional model
