一种基于证据理论的多传感器目标识别方法

Multi-Sensor Target Recognition Method based on Evidence Theory

证据理论是高层数据融合中的一种重要方法,因其能够很好地处理不确定性问题,近些年来被广泛应用于决策判断、目标识别等数据融合领域。随着人们认知能力的不断提高和科技的不断发展,以及感知信息识别框架的不断完善,如何给感知信息赋予一个合理而准确的基本概率值成了证据理论发展的一大研究重点。经过学习研究,提出了通过计算感知信息与先验知识之间的距离来生成基本概率赋值函数,再对其进行证据融合的方法。最后经过数据验证,发现融合结果的准确度较高,符合预期。

Nowadays, evidence theory, as an important method for high-level data fusion, is widely adopt- ed in the fields of decision-making judgment and targets tracking for its fairly dealing with uncertain is- sues. With the increasing of people＇s awareness, the development of science and technology and the con- stant perfection of perceptive-information recognition framework, how to assign perceptive information a reasonably accurate value becomes a major research topic in the development of evidence theory. This pa- per puts forward a method to generate basic belief assignment function by calculating the distance from pri- or knowledge to perception information, and then to fuse them through the evidence theory. Experiment in- dicates that the fusion result has a good veracity and is accordant with the expectation.