摘要
针对Dempster-Shafer理论(DST)及Dezert-Smarandache理论(DSmT)难以处理不确定信息的问题,定义了辨识框架中的不确定因子,提出了2种自适应通用分配法则(AUPR).并提出了证据理论的广义融合框架,并在此基础上构建了广义证据推理机.以Pioneer 2-DXe机器人为实验平台,绘制了实验场景的信度分布图.实验结果验证了所提方法的有效性和实用性,为构建统一的信息融合框架提供了有力的依据.
In order to solve the problem of the Dempster-Shafer theory（DST） and Dezert-Smarandache theory（DSmT） both being unable to deal with uncertain information,the uncertainty mass in the frame was defined and two kinds of adaptive universal proportional redistribution rules（AUPR） were proposed.Next,a general evidence reasoning fusion structure was proposed based on the general evidence with which the reasoning machine was built.Lastly,the pioneer 2-DXe mobile robot was used to build the belief distribution maps of various environments.The experimental results verify the validity and the practicality of the proposed methods.They also supply powerful theoretical evidence for constructing a uniform information fusion frame.
出处
《智能系统学报》
2010年第6期487-491,共5页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(60675028)
关键词
证据推理
融合框架
地图构建
信息融合
evidence reasoning
fusion frame
map building
information fusion
