摘要
DS理论(DST)在低冲突条件下融合精度高于DSm理论(DSmT),而DSmT在高冲突条件下融合精度高于DST。为进一步提高融合精度,结合DST和DSmT的优点,提出一种基于DST和PCR2的自适应信息融合算法。对DSmT框架下的分配法则进行分析,选择融合精度较高且时间复杂度较低的PCR2分配法则。采用PCR2和DST进行自适应融合,即在冲突率高的情况下使用PCR2,而在冲突率低的情况下使用DST。算例结果表明,该算法能够进一步提高融合结果的准确程度,时间复杂度介于DST和PCR2规则之间。
Dempster Sharer Theory(DST) fusion accuracy is higher than Dezert Smarandache Theory(DSmT) under low-conflict conditions, and DSmT fusion accuracy is better than DST under high-conflict conditions. In order to obtain more accurate fusion results, through combining advantages of DST and DSmT, this paper proposes an adaptive information fusion algorithm based on DST and PCR2. The allocation rules under DSmT framework are analyzed, and the rule of PCR2 is selected because of its higher accuracy and lower complexity. Adaptive fusion algorithm is given using PCR2 and DST, PCR2 is employed when there are higher conflicts, and DST is employed when there are lower conflicts. Numerical results show that this algorithm obtains better fusion results and its complexity is between DST and PCR2.
出处
《计算机工程》
CAS
CSCD
2013年第5期69-72,77,共5页
Computer Engineering
基金
国家自然科学基金资助项目(61201118)
陕西省教育厅科研计划基金资助项目(12JK0529)
西安工程大学博士科研启动基金资助项目(BS1111)
关键词
DS理论
DSm理论
PCR2规则
自适应信息融合
证据理论
Dempster Shafer Theory(DST)
Dezert Smarandache Theory(DSmT)
PCR2 rule
adaptive information fusion
evidence theory