华中科技大学学报（自然科学版）2025，Vol.53Issue(5)：44-51,8.DOI:10.13245/j.hust.250796
基于数据删除与合并的相似轨迹查询性能优化
Performance optimization of similar trajectory query based on data deletion and compaction
摘要
Abstract
To address the issue of the multistage task parallelization method for similar trajectory query(MPST)generating a large amount of useless and redundant data during vehicle trajectory queries based on trajectory similarity,leading to degraded system performance,a method for eliminating useless data and compacting redundant data based on ring-domain calculation was proposed.Specifically,a ring-domain structure to distinguish the attributes of point-accompanying vehicle pairs—either incidental or effective accompanying attributes was employed by this method.By removing incidental accompanying vehicle pairs(i.e.useless data)and compacting effective accompanying vehicle pairs(i.e.redundant data),the size of intermediate results was reduced,thereby decreasing system I/O overhead and shortening response time.To validate the effectiveness of this method,a prototype system called eliminating and merging strategy based on ring for similar trajectory query(EMSR)was implemented.Experimental results show that compared with MPST,the intermediate result set generated by EMSR is only 5.7%～11.2%of that of MPST,and the system response time is reduced to 30.2%～57.7%of that of MPST.关键词
相似轨迹/伴随车辆/无用数据删除/冗余数据合并/车牌识别数据Key words
similar trajectories/companion vehicles/useless data deletion/redundant data compaction/automatic number plate recognition分类
信息技术与安全科学引用本文复制引用
潘玉彪,林宜,张惠臻..基于数据删除与合并的相似轨迹查询性能优化[J].华中科技大学学报（自然科学版）,2025,53(5):44-51,8.基金项目
国家自然科学基金资助项目(61802133) （61802133）
福建省自然科学基金资助项目(2021J01319). （2021J01319）
