摘要
受生物体神经内分泌系统调节机制的启发,提出一种神经内分泌计算模型.该模型中,内分泌系统能够对神经系统的学习与记忆行为进行反馈调控,使自主体及时调整行为,从而提高其学习和适应未知环境的能力.为了验证模型及算法的有效性,将其应用于机器人导航避障仿真实验,并与离散Q学习方法对比,结果表明该模型是有效的.
Inspired by the regulation mechanism of biological neuro-endocrine system,a neuron-endocrine computation model is proposed.In the model,the endocrine system regulates the learning and memorization behaviors of the neural system by feedback control,which could adjust the agent＇s behavior in time and improve its ability to learn and adapt to the unknown environment.To verify the validity of the model,we performed a simulation experiment,in which the proposed model and Q-learning with discrete states were utilized respectively to help a robot learn the ability of obstacle avoidance.Experimental results show that the proposed model is more effective than the Q-learning with discrete states.
出处
《小型微型计算机系统》
CSCD
北大核心
2010年第9期1910-1913,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60875027)资助
关键词
内分泌系统
神经网络
神经内分泌系统
避障
endocrine system
neural network
neuro-endocrine system
obstacle avoidance
