ICS25.040.99 CCSl66 SAIAS 上海市人工智能行业协会团体标准 T/SAIAS018—2025 具身智能语料库建设导则 ConstructionGuidelinesforCorpusofEmbodiedIntelligence 2025-04-30发布 2025-05-01实施 上海市人工智能行业协会  发布 全国团体标准信息平台 全国团体标准信息平台 全国团体标准信息平台 全国团体标准信息平台 T/SAIAS018—2025 I目次 前 言............................................................................IV 引 言............................................................................VI 1范围.................................................................................1 2规范性引用文件.......................................................................1 3术语和定义...........................................................................1 4缩略语...............................................................................2 5具身智能数据.........................................................................2 5.1一般原则.........................................................................2 5.2数据集...........................................................................2 5.2.1数据片段.....................................................................2 5.2.2时间切片.....................................................................3 5.3数据资源.........................................................................3 5.3.1数据资源类型.................................................................3 5.3.2数据信息存储格式.............................................................3 6具身智能数据采集.....................................................................4 6.1真实数据采集.....................................................................4 6.1.1采集方式.....................................................................4 6.1.2采集本体.....................................................................5 6.2合成数据采集.....................................................................5 6.2.1采集方式.....................................................................5 6.2.2合成数据有效性考虑因素.......................................................5 6.2.3合成数据生成方式.............................................................5 6.2.4合成数据用途.................................................................6 6.3采集数据提交.....................................................................6 6.3.1数据文件标识.................................................................6 6.3.2数据提交方式.................................................................6 6.4数据采集场景与任务体系...........................................................7 6.4.1场景.........................................................................7 6.4.2任务.........................................................................7 6.4.3原子技能.....................................................................7 6.5数据采集操作规范.................................................................7 6.5.1采集环境要求.................................................................7 6.5.2采集设备选型.................................................................7 6.5.3采集人员培训.................................................................7 7具身智能语料生产.....................................................................8 7.1采集数据完整性检查...............................................................8 7.2数据资源清洗过程.................................................................8 全国团体标准信息平台 全国团体标准信息平台 T/SAIAS018—2025 II7.2.1主要流程.....................................................................8 7.2.2数据规范管理.................................................................8 7.2.3数据整理分类.................................................................8 7.2.4数据脱密脱敏.................................................................8 7.2.5数据去噪.....................................................................9 7.2.6数据去重.....................................................................9 7.2.7数据存储与备份...............................................................9 7.3数据资源标注过程.................................................................9 7.3.1基本要求.....................................................................9 7.3.2标注任务制定...........................................................