Abstract: The purpose of this study is to present a method of driving rule acquisition using variable precision rough set theory.This method can solve the problem of rule acquisition under condition of imprecise and inconsistent information on account of kind overlapping.Rough set theory has been aiming at data analysis problems involving uncertain or imprecise information，its main idea being to acquire concept’s classifying rules through knowledge reduction maintaining classify ability.As the extension of rough set，variable precision rough set reflects the difference of degrees of overlap of kinds and sets.This paper presents a method of rule acquisition based on variable precision rough set theory，explains its application method using an example of multi-information fusion during driving，and proves its validity.

Key words: rough set, variable precision rough set, rule acquisition, multi-information fusion

摘要： 研究了应用可变精度粗糙集获取驾驶规则的方法。该方法的特点是可以处理由于类重叠引起的样本信息不精确、不一致情况下的规则获取。粗糙集理论一直用于研究不确定或不精确信息的数据分析问题，其主要思想是在保持分类能力不变的前提下，通过知识约简，导出概念的分类规则。而可变精度粗糙集作为对经典粗糙集理论的扩展，体现了等价类与集合的重叠度程度上的差别。给出了基于可变精度粗糙集的获取规则的方法，以驾驶过程的多源信息融合实例说明其使用方法，并验证了其有效性。

关键词: 粗糙集, 可变精度粗糙集, 规则获取, 多源信息融合