基于BP网络的机器人多传感器手爪特征级数据融合  

BP Network Based Feature-level Data Fusion for Robotic Multi-sensor Gripper

作　　者：徐科军[1,2] 童利标[1] 徐科军[1,2] 梅涛 

机构地区：[1]合肥工业大学,合肥230009 [2]中国科学院合肥智能机械研究所,合肥230031

出　　处：《仪器仪表学报》2002年第5期526-529,533,共5页Chinese Journal of Scientific Instrument

基　　金：航空863项目资助。

摘　　要：机器人手爪上装有多种传感器。为了保证手爪能安全、可靠地抓取工件,必须对这些传感器的数据进行融合。本文介绍手爪上传感器的配置情况,进行手爪抓取工件实验,将BP神经元网络应用于传感器的特征级融合,得到手爪抓取工件的准确信息。Several kinds of sensors are installed in the robotic gripper. In order to grasp objects safely and reliably, the information fusion should be carried out for the output data of multi-sensors. In this paper,the configuration of sensors in the gripper is introduced,the grasping object experiments with a gripper are made, the results of the feature-level fusion are presented by BP neural network,the accurate information of grasping object with a gripper are obtained.

关 键 词：机器人手爪 多传感器数据融合 BP神经元网络 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]