摘要
从信度赋值技术的角度,提出了统一融合框架(UFA)的理论机制,它是基于证据支持贴近度(ESMS)前置过滤器耦合包含冲突分配器在内的融合机,在给出一个ESMS函数的同时,也考虑了不同信息源(可靠、非可靠、不完全、冗余及互补)信度赋值中心,克服了错误、虚假和不一致等信息对融合机的干扰,充分发挥了融合机优势,减少了融合计算的复杂度,大大提高了融合效果.将统一融合框架应用于Pioneer II移动机器人地图创建方面,通过仿真分析,并与单一融合机的仿真结果作了对比,充分说明了统一融合框架的优越性.
Unified fusion architecture （UFA） is proposed in terms Of belief assignment technology, which refers to pre-processor of evidence supporting measure of similarity （ESMS） filter with the fusion machine including conflict redistributor. Here an ESMS function is given, and barycentres of belief masses from different kinds of information source （reliable, unreliable, incomplete, paraconsistent and mutually complementary） are considered, so that we overcome some disturbing influences on the fusion machine from mistaken, illusive and inconsistent information, exert the advantage of fusion machine adequately, reduce the complexity of fusion computation, and improve the fusion effect greatly. At last, we apply unified fusion architecture to map building of Pioneer II mobile robot. By simulation analysis and even comparison from the simulation result without the pre-processor, it shows the advantage of unified fusion architecture adequately.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2007年第1期46-49,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(69585003)
关键词
比例冲突重新分析
信息融合
统一融合框架
支持贴近度过滤器
proortional conflict redistsibution （PCR）
（UFA）
ESMS filter information fusion
unified fusion architecture
