摘要
人工情感在机器人的研究中至关重要,文中简要概括了当前人工情感的应用;我们在借鉴生物系统控制理论的基础上,融入了进化控制的思想,设计了一种基于人工情感的控制体系结构,在此结构中包含有基于蚁群算法的进化控制系统、神经和人工情感控制系统;机器人通过神经系统接受环境信息并进行行为决策,行为决策的效果通过情感学习模型进行反馈;情感学习模型根据机器人的内、外环境状态,产生情感因子(即生物激素),再由情感因子来调节神经系统的记忆和行为决策,最后神经系统的记忆与行为模块又由进化系统得以继承;该控制结构加强了机器人在动态环境中的学习和自适应能力;为了验证该控制结构的有效性,文章做了仿真实验;仿真结果也表明机器人具有很强的学习和自适应能力。
Artificial emotion is essential to study in robotics. Some typical implements of artificial emotion are briefly summarized in this paper. Motivated by the biological system control mechanism, a control architecture based on artificial emotion is proposed, which includes evolution control system based on Ant Algorithm, neural and artificial emotion control system. Neural system receives environmental information and make decisions. The result of decision--making is fed back by emotion--learning model. The emotion--learning model produces e-motional factors (hormones) based on inner and outer conditions of the robot, then uses these factors to regulate neural system, at last the memorization and behavior mode of neural system is exported to genetic environment. The control architecture enhances learning and adaptive capacity of robot in dynamic environment. In order to prove the validity of the control architecture, we demonstrated it by thesimulation. Experimental evidence of successful controlling is also presented.
出处
《计算机测量与控制》
CSCD
2007年第10期1388-1391,共4页
Computer Measurement &Control
基金
湖北师范学院研究生启动基金
关键词
神经网络
蚁群算法
人工情感
行为决策
机器人
neural networks Ant algorithms artificial emotions decision--makings robot