Abstract: As an important inference theory of uncertainty, D-S evidence theory provides a good solution to deal with the ambiguity and uncertainty of sensor information. However, how to construct basic probability assignment function（mass function） is still an open issue. To solve this problem, this paper proposes a Gaussian fuzzy membership function based on fuzzy theory to obtain the reliability of each sensor, and calculates the mutual supportability of multiple sensors. Then the reliability and supportability are changed to mass function. Finally, multi-sensor information is combined by using evidence theory. Simulation results show that this method can improve the accuracy and reliability of the recognition.

Key words: evidence theory, information fusion, fuzzy theory, membership function

摘要： D-S证据理论作为一种重要的不确定性推理理论，为处理传感器信息的模糊性及不确定性提供了很好的解决方法。但各个证据中的基本概率分配函数（mass函数）如何生成，仍是人们需要解决的问题。针对这一问题，提出了一种基于模糊理论中的高斯隶属度函数来得到传感器提供信息的可信度，计算了各个传感器之间的相互支持度；将各传感器的可信度和支持度转化成mass函数；利用证据理论对多传感器信息进行融合。仿真试验表明该方法能够有效提高识别的准确性和可靠性。

关键词: 证据理论, 信息融合, 模糊理论, 隶属函数