摘要
针对鼠类行为的不可控性、不可预测性和不可重现性等问题,提出一种可控的、可预测的和可重现的智能交互控制系统.通过控制自主设计的双目机器鼠输出不同的行为状态(恐惧、中立、友好)来刺激实验生物鼠,研究生物鼠的回应行为.该系统采用双目视觉算法,可以有效获取实验鼠的三维立体信息,结合一套高效的支持向量机图像分类处理算法,可以准确地分析实验鼠的行为状态,然后通过调整机器鼠的行为来改变实验鼠的行为状态,实现智能交互.机器鼠与生物鼠交互实验证明:该系统可以实时捕获生物鼠的行为状态,并且通过控制机器鼠的输出行为,定向改变生物鼠的行为.该系统可以为生物学家、神经科学家和脑科学研究者等提供研究平台和辅助研究工具.
To address the uncontrollability,unpredictability and non reproducibility of rat's behavior,an intelligent interactive control system was proposed.Different behaviors(stressful,neutral,friendly),which aims to stimulate a laboratory rat were generated by controlling a self-designed roboic rat with binocular vision.The response of a rat to a robotic rat was studied.The three-dimensional information of rats were effectively obtained based on binocular vision system.The experimental rat's behavioral characteristics was accurately analyzed by combining with an efficient image classification processing algorithm,and the behaviors of rats were modulated by controlling the behavior of the robotic rat,so as to achieve intelligent interaction.The social interaction experiments between the laboratory rats and the robotic rat show that the system is able to capture rat's behavioral characteristics in real time and direct rat's behavior by controlling the behavior of robotic rats.The system provides biologists,neurologists and brain scientists with a research platform and an assistant research tool.
作者
王宏民
王子荣
宋永博
WANG Hongmin;WANG Zirong;SONG Yongbo(School of Electric and Engineering,Harbin University of Science and Technology,Harbin 150080,China;Key Laboratory of Complex System and Intelligence Science,Institute of Automation,China Academy of Sciences,Beijing 100190,China)
出处
《北京工业大学学报》
CAS
CSCD
北大核心
2018年第1期64-72,共9页
Journal of Beijing University of Technology
基金
黑龙江省自然科学基金资助项目(F201310)
关键词
机器人
控制系统
智能交互
机器鼠
双目视觉
图像处理
robot
control system
intelligent interaction
biorobotic rat
binocular vision
imageprocessing