Information-Based Task-Level Planning

这项研究启动奖是为机器人零件装配的理论和实验工作。理论研究将确定简单操作任务的信息不变量：感知、行动和预测之间固有的组装时间权衡。然后，厄德曼博士将在一个机器人零件定位系统上测试他的理论，该系统将沿着信息不变的边界分解感知和操作任务。这项研究可以改善制造设计以及可靠和高效的零件装配。

This Research Initiation Award is for theoretical and experimental work in robotic parts assembly. The theoretical research will identify information invariants for simple manipulation tasks: inherent assembly-time tradeoffs between sensing, action, and prediction. Dr. Erdmann will then test his theory on a robotic parts-orienting system, with sensing and manipulatory tasks decomposed along information-invariant boundaries. This research could lead to improved manufacturing design as well as reliable and efficient parts assembly.