Physics-based manipulation planning and control for bin-picking applications

该项目旨在开发规划和控制算法，以及机器人系统，以执行从杂乱的多物体垃圾箱中挑选物品的操作。虽然基于物理的计划和控制已经应用于平面上物体（如架子或桌子）的简单操作领域，但垂直进入盒子和箱子（例如在仓库设置中）的问题尚未研究。该项目将研究轨迹优化，以及基于采样的方法，以规划和控制该领域。对计算效率和对不确定性的鲁棒性的主要挑战进行了评估。开发的算法将与许多工业应用相关，特别是仓库订单交付拣选和包装。

This project aims to develop planning and control algorithms, as well as robotic systems, to perform picking operations of items out of cluttered multi-object bins. While physics-based planning and control have been applied to the simpler domain of manipulation of objects on a flat surface such as a shelf or a table, the problem of reaching vertically into boxes and bins, e.g. in a warehouse setting, has not been investigated yet. This project will investigate trajectory optimisation, as well as sampling-based, approaches to planning and control in this domain. The main challenges of computationally efficiency and robustness to uncertainty will be evaluated. Developed algorithms will be relevant to many industrial applications, particularly warehouse order delivery picking and packing.