摘要
针对自动泊车系统泊车精度评价过程中遇到的试验数据样本量不大、系统误差无法忽略且难以量化等困难,采用Bootstrap方法理论构建小样本条件下泊车精度综合评价方法。通过实车试验收集泊车精度偏差样本数据,充分利用有限试验数据样本信息,开展重抽样,生成多组同分布的泊车偏差数据,在90%置信度下对泊车精度评价结果进行检验,结果准确可信。
A comprehensive evaluation method on accuracy was built for the auto-parking system based on Bootstrap. The method could solve the problem of small sample size, assignable system error which was also difficult to be quantified.The parking sample data of precision was collected from practical experiments,the limited sample data was used fully to resample the original samples, then the sample data groups distributed identically were generated. Finally, the results are checked in 90% confidence level and prove that the evaluation method for auto-parking system based on Bootstrap is convenient and effective.
出处
《机械设计》
CSCD
北大核心
2014年第12期71-74,共4页
Journal of Machine Design
基金
江苏省产学研联合创新资助项目(BY2012173)
江苏省高校自然科学基金资助项目(12KJD580002)
交通部信息化重点资助项目(2013-364-863-900)
中国博士后基金资助项目(2013M531283)
江苏省博士后基金资助项目(1201037C)
江苏大学高级人才基金资助项目(1281170019)
江苏省研究创新计划资助项目(KYLX_1057)
