项目名称: 基于限制释放算法驱动的发展型机器人及其书写动作控制
项目编号: No.61203336
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 自动化学科
项目作者: 晁飞
作者单位: 厦门大学
项目金额: 22万元
中文摘要: 发展型机器人为智能机器人的研究新热点,其强调机器人通过模仿人类发育模式来实现对自己的感觉运动机构和外部世界的认知。本课题以发展型机器人和发展心理学的研究为基础,参考婴儿发育模式中的特征,采用并改进"限制释放"这一全新的发展型学习算法来分析、构建机器人的发展过程,用发育机制改进的梯级关联神经网络搭建模仿婴儿大脑的模型为方法,解决发展型机器人能够通过不断的发展达到学习模仿人书写毛笔字的能力这一问题。课题研究将通过机器人在人书写动作模仿的演示,来检验和展示上述模型及其关键技术实现的效果,以阐明机器人高级的自主认知与学习能力是可以通过发展的方法来实现。课题的研究意义体现在为把发展心理学与发展神经学理论转化成发展型机器人的构建方法开拓一种可行的途径,这对丰富智能机器人构建方法、促进国内发展型机器人的研究都有着重要的学术意义。也对在恶劣或极端环境下机器人通过模仿人的行为完成复杂动作的研究具有应用意义。
中文关键词: 发展型机器人;机器人书写动作控制;发展型学习算法;人-机器人交互;
英文摘要: Developmental robotics is a new research field of intelligent robots. It focuses on constructing its sensory motor mechanism and its internal representation of the world by mimicking human infant development model. This project relies on studies on developmental psychology and developmental neuroscience, refers to features of human infant development, and adopts the enhanced "Constraint Lifting" algorithm to analyze and build robotic developmental procedure. This project also attempts to apply developmental mechanism to implement a brain-like model by using improved Cascade-correlation neural networks. The objective of the project is to solve the problem that developmental robots can autonomously learn to write Chinese characters. The project demonstrates that high level cognition and learning abilities can be achieved by developmental approaches. Its significances contain that the project effectively merges developmental psychology and neuroscience into robotics; and the project is very useful for increasing robotic construction solutions, and important for enhancing research on developmental robotics in China; and it is also crucial for that robots are capable of working under complex and extreme environments.
英文关键词: Developmental robotics;control of robotic writing motion;developmental learning algorithm;human-robot interactions;