摘要
基于神经网络和模糊积分的理论提出了一种多传感器信息的空间和时间两级数据融合结构,并将其用于机器人的物体识别,提高了系统的识别效率,增强了系统的可靠性。
In this paper, based on the theory of neural networks and fuzzy integral, a spatial-temporal two-layer architecture for sensor fusion is presented. It can improve the robustness and reliability. The simulation with the robotic multisensor object identification is also presented and discussed.
出处
《电机与控制学报》
EI
CSCD
1998年第2期112-114,118,共4页
Electric Machines and Control
基金
国防科工委预研基金
关键词
传感器融合
神经网络
物体识别
机器人
sensor fusion
neural network
fuzzy integral
object identification