Think inside the box : 3D bin packing visualized with spatial AR andautomated depth feedback

Abstract: In modern day logistics, small and medium sized consumer goods are often manually handled and packaged in fulfillment centres and warehouses. When time efficiency is prioritised, there exists the risk to pick too big a package and thus waste space which can naturally lead to unwanted consequences such as unnecessary transports being booked and excess package material being used. This in turn leads to cost inefficiency and increased environmental impact. One novel way of tackling this issue is to use algorithms made to solve optimization problems such as theff3D bin packing problem to pre-calculate which package to use when assembling an order. However, as fill rates improve, it may become difficult for warehouse workers to fit all items into the selected package, especially when dealing with a larger number of items. To help with this, modern techniques such as spatial augmented reality can be used to visualize the packing process in real time, showing workers which item to place next, and where. Recent studies have shown improved packing speeds using this approach, however further improvements could potentially optimize the usability even further. In this study, a novel spatial AR application was used in conjunction with a depth camera to not only indicate where to place items, but to progress the instructions automatically only if they were placed correctly. The aid-tool was tested on 26 participants and showed significant improvements by a factor of 3.29 for a fill rate of 90%, slight but insignificant improvements by a factor of 1.18 for a fill rate of 85% and performed worse by a factor of 0.71 for a fill rate of 81% when compared against unassisted packing. The study concluded that the chosen spatial AR tool showed promise for fill rates of 90% and above, but the results should be treated with caution as the sample size and methodological flaws severely limited the scope of the findings.