Delay Impacts on EEG-Based Determination of the Human Visual Interface QoE for Virtual and Augmented Realities

The emergence of 5G services includes those requiring extremely low-latency for command and control application scenarios in the low millisecond range. With a significant human-machine interaction component, the Tactile Internet will require that the experiences based on the human visual interface can be dynamically adjusted within similar time frames. In this paper, we evaluate the impact that different delays currently attainable with commercial hardware would have on predicting the Quality of Experience (QoE) with immersive images. Specifically, we employ electroencephalography (EEG) data to predict how future subjects would determine the media quality in a Passive Human In-the-Loop (PHIL) scenario. This initial extension of our prior work focuses specifically on the delay in the gathering and processing of data and presents a first foray of bringing the passive human-in-the-loop QoE adjustment approach to the Tactile Internet. We find that there is limited value in increasing the delays of two different approaches to predicting the QoE. We additionally note that current approaches to predicting a user's QoE based on other users' EEG patterns exhibit only limited prediction accuracy.