How wrong can you be: Perception of static orientation errors in mixed reality

Abstract

Tracking technologies are becoming an affordable commodity due to the wide use in mobile devices today. However, all tracking technologies available in commodity hardware is error prone due to problems such as drift, latency and jitter. The current understanding of human perception of static tracking errors is limited. This information about human perception might be useful in designing tracking systems for the display of AR and VR scenarios on commodity hardware. In this paper we present the findings of a study on the human perception of static orientation errors in a tracking system, using two different setups leveraging a handheld viewfinder: a classical augmented scenario and an indirect augmented one. By categorizing static orientation errors by scenario and local orientation axis, new insights into the users' ability to register orientational errors in the system are found. Our results show that users are much more aware of errors in classical AR scenarios in comparison to indirect AR scenarios. For both scenarios, the users registered roll orientation errors differently from both pitch and yaw orientation errors, and pitch and yaw perception is highly dependent on the scenario. However, the users performance ranking for orientational errors in AR scenarios was unexpected.

Registration errors in beacon-based navigation guidance systems: Influences on path efficiency and user reliance

“…Particularly at farther distances, angular errors exceeding 15-20 degrees are very apparent and likely noticed by participants. This is similar to perceptible registration errors perceived and explicitly acknowledged by participants using other augmented reality systems (Madsen and Stenholt, 2014).…”

“…For instance, magnetic compasses can vary widely (+/-5°) in accuracy due to variability in the Earth's magnetic field, and Global Positioning Systems (GPS) can produce dynamic orientation errors on the order of +/-10° (Azuma, 1997). A recent study demonstrated that even a stationary gyroscope (in an iPad 3) demonstrates yaw drift of +/-10˚ over the course of 12-18 hours (Madsen and Stenholt, 2014). Of course, certain environments such as those characterized by tall buildings (i.e., urban canyons) or mountains, can result in even more severe positioning and relative direction errors (Duncan et al, 2013;Miura et al, 2015).…”