Analysing the use of 5G for vehicular AR applications

Abstract

Augmented Reality is a technique for superimposing video cues onto a real-world display to allow a better interpretation of the real-world view. Mobile AR uses mobile devices such as smartphones or tablets, to view the real world through their cameras and superimpose cues onto the device's video display. When used in vehicles, AR can help drivers navigate better and more safely through their surroundings. The purpose of this dissertation is to analyse how the 5th Generation in mobile telecommunications (5G), can help deploy useful Mobile AR applications by exploiting the high throughput, Ultra-Reliable Low Latency Communications (URLLC), to deliver video from the road to the MAR device in real-time so that it can be overlaid over the view seen through the vehicle's windshield. A MAR system is developed with a 5G smartphone acting like a video camera that can capture video of oncoming traffic and relay it to a second 5G smartphone in the vehicle operating as a MAR display. Two applications are specified, designed, implemented, and tested; a Virtual Mirror and an Obstructed View Eliminator. The video is transmitted to the MAR device in real-time over 5G, 4G-LTE and Wi-Fi to allow the comparison of end-to-end Latency, which is the sum of network latency and video encoding and decoding latency. Objective testing shows that the 5G achieves the lowest network latency, below 10ms, followed by Wi-Fi and 4G LTE. The video end-to-end delay is also lowest with 5G and averages 200ms, with the lowest latency of 128ms. Subjective testing was also conducted to assess the suitability and acceptability of the VM and OVE applications. The majority of respondents rated the systems positively and would use the system, although some expressed some concern about possible distractions during driving.