“复杂环境下面向多任务的协同、柔顺、精准操控”关键考核指标第三方检测数据

针对探测环境复杂、任务多样、先验知识欠缺、资源严重受限等问题，开展高效操控学习、知识传递与协同柔顺操控研究，突破多任务多约束高可信自主规划与决策、不确定因素下感知执行交互的精准操控学习训练与模型、严苛环境下高效协同以及知识传递关键技术，实现地外未知环境中的操作经验积累与操作性能优化。本数据集基于项目总体搭建的无人系统进行火星模拟场的操控实验采集，主要记录了无人系统在高价值目标样品采集、自然柔顺精准操控、多任务规划时间以及决策正确率的方面的测试数据。

Aiming at the challenges including complex detection environments, diverse tasks, insufficient prior knowledge, and severely limited resources, this study conducts research on efficient manipulation learning, knowledge transfer, and collaborative compliant manipulation, breaks through key technologies such as high-reliability autonomous planning and decision-making under multi-task and multi-constraint scenarios, precise manipulation learning, training and models for perception-execution-interaction under uncertain factors, efficient collaboration in harsh environments, and knowledge transfer, and achieves the accumulation of operational experience and optimization of operational performance in unknown extraterrestrial environments. This dataset is collected from manipulation experiments carried out in a Mars simulation field using the unmanned system developed under the overall project framework, and mainly records test data of the unmanned system in terms of high-value target sample collection, natural compliant precise manipulation, multi-task planning duration, and decision-making accuracy.

搜集汇总

背景与挑战

背景概述

该数据集记录了无人系统在火星模拟场中的操控实验数据，重点测试了高价值目标样品采集、自然柔顺精准操控、多任务规划时间及决策正确率等关键指标。数据由上海交通大学提供，属于人工智能领域，数据量为2.64MB，包含2个文件。