Tracking Support for Collaborative Warehousing using Head-mounted Augmented Reality

Abstract

In this thesis, we show that the conventional workflows for tracking objects in a warehouse can be enhanced by a new Augmented Reality (AR) service. In particular, we have designed a Unity implementation of an AR toolkit for the Microsoft HoloLens based on a "indoor greenhouse" scenario, which enables the physical locations of plants (together with plant data such as nutritional status) to be tracked as they are moved by greenhouse workers. As part of this case study, we found that a considerable latency occurs when sharing the physical positions of the plant tray between multiple users in a collaborative scenario. To deal with this, we have improved Microsoft's Local Anchor Transfer method which only enables one "spatial anchor" (corresponding to the information of the physical location of an object) to be transferred for one object. Subsequently, we have designed a new One-For-All-Shared-Experiences (OFALL-SE) method to enable many objects to be tracked by one Spatial Anchor. Our results show that use of the OFALL-SE method means that the transmission latencies can be greatly reduced.

Although our OFALL-SE method can implement a near-real-time transmission of the spatial anchor when tracking multiple objects, its associated spatial limitations mean that it can only track objects through small regions of space. Following the recommendations of Microsoft, their Azure Spatial Anchor (ASA) service can overcome this limitation and used in a large space, but it requires a new toolkit which is independent with our OFALL-SE technique. ASA also requires external implementations by developers and it also lacks security on the user's anchor information. Thus, we have designed a new technique based on our OFALL-SE technique - Spatial Anchor Based Indoor Asset Tracking (SABIAT) - that achieves similar system features as ASA service such as multiple spatial anchors management over a large area. SABIAT can be implemented as a standalone application or integrated into other AR applications to provide location-based services such as our OFALL-SE method. With this new technique, We also created a standalone AR application - AR-IPS, which show that objects can hop between different spatial anchors, and the historical positions of multiple objects can be tracked, visualized and recorded. Show more