Computer vision based augmented reality for guiding and evaluating assembly sequences

Abstract

Summary form only given. Augmented reality (AR) provides an intuitive interface to enhance the user's understanding of a scene. We consider the problem of scene augmentation in the context of assembly of a mechanical object. Concepts from robot assembly planning are used to develop a systematic framework for presenting augmentation stimuli for this assembly domain. An interactive evaluation tool is developed, which uses augmentation schemes for visualizing and evaluating assembly sequences. This system also guides the user step-by-step through an assembly sequence. Computer vision provides the sensing mechanism necessary to interpret the assembly scene. The goal of this system is to help evaluate the feasibility and efficiency of a particular sequence to assemble a mechanical object from its components. This is done by guiding the operator through each step in the sequence. The augmentation is provided with the help of a see-through head-mounted display that superimposes 3D graphics over the assembly scene and on nearby computer monitors. We incorporate these ideas into the design of an integrated system that we call AREAS (Augmented Reality System for Evaluating Assembly Sequences) and explore its use for evaluating assembly sequences using the concept of mixed prototyping.