模糊—证据理论信息融合算法

Fuzzy-Evidence Theory Information Fusion Algorithm

研究多传感器信息融合优化算法。针对多传感器之间相互支持程度,传统计算信息融合的算法存在绝对化的问题,提出了一种证据理论和模糊集合的信息融合方法,首先利用相关性函数定义不确定信息的模糊支持区间和模糊支持概率,然后由隶属函数得到各个传感器提供信息的可信度,再将支持度和可信度转化为基本概率分配函数,最后进行D-S证据合成。上述方法摒弃了以往支持概率选取的绝对化,在概率融合意义下定义模糊置信距离测度和一种新的支持概率。仿真结果表明,改进方法获得的结果具有更高的精度和可信度。

Optimization of multi - sensor information fusion was studied in this paper. 1Regarding the low mutual - support among sensors and the problem of absolutization in traditional information fusion algorithms, an algorithm based on evidence theory and fuzzy set was proposed. First, the correlation functions were used to define fuzzy sup- port intervals and probabilities; then the membership functions were used to calculate information credibility for each sensor, and the support values and credibility were then transformed to basic probability distribution functions. Final- ly the D - S evidences were combined. This algorithm defined fuzzy confidence distance measurement and a new type of support probabilities based on probability fusion, rather than absolutely selected support probabilities like previous methods. Simulation results show that this improved algorithm has higher precision and reliability.