Uppsats

QoE Assessment of Cloud-based Social Extended Reality Applications over Heterogeneous Access Networks

M1-uppsats

Luleå tekniska universitet/Institutionen för system- och rymdteknik

Publicerad: 2025

Språk: Engelska

Sammanfattning

XR applications and devices have gained popularity lately, seeing increasing growth with everypassing year from both private and public actors. This thesis will explore the efects that constrainedmobile network conditions, including RTT, PL, and Jitter, impose on social extended reality applica-tions over heterogeneous access networks(such as 4G, 5G, and WiFi), and their combined efect on theuser’s perceived quality of experience. QoE and MOS results from subjective tests will be presentedand discussed at the end. Additionally, recommendations for what constraints should be adhered tofor the best possible user experience.An experimental test bed was constructed to conduct the subjective user tests and Resonite, a so-cial VR application was chosen from within which the tests would be conducted as it allowed for thecreation and management of a headless server. The test bed consisted of two Oculus Quest 3 VRheadsets, ALTRUIST software(to impose various network conditions on the participant’s device andmanage the user data), server (to host the Resonite private virtual room). In total, 20 conditions wereimposed on the participant’s device throughout each test and subjective data was collected utilizingquestionnaires prompted to users. 33 users were gathered for this study and using the Calendly appli-cation for scheduling participant appointments.The results showed that audio quality was signifcantly afected by jitter, even at small values ofjitter=1. This poses major challenges as it was shown in [1] that the average jitter over wireless con-nections(4G, 5G-NSA, and 5G-SA) exceed 2.5ms at all times, even reaching peaks of 7.5ms averagejitter. Otherwise, audio quality preserves its quality for RTT up to 77 ms and PL up to 6%. QoE wasshown to not be afected in any signifcant way by RTT up to 77ms, although surpassing RTT=77msand PL=2% should be avoided. Jitter afects QoE signifcantly as well, where jitter=1 surpassingRTT=52ms results in signifcant degradation in overall user experience. Any jitter values beyond 3standard deviation for RTT=27ms and beyond should be avoided. The video quality sees consider-able video degradation when surpassing RTT=27ms and RTT=27 combined with jitter=6 also seessignifcant degradation

Information

Författare
Gavrell, Justin
Lärosäte / institution
Luleå tekniska universitet/Institutionen för system- och rymdteknik
Publiceringsdatum
2025
Uppsatstyp
M1-uppsats
Språk
Engelska

Utforska vidare

Liknande uppsatser

Uppsatser med liknande ämnen och nyckelord.