Hoppa till huvudinnehåll
RISE logo

Master Thesis: Low latency Render Streaming on 5G/Edge for High-Definition Web AR/VR applications

With the emergence of Edge Computing and 5G networks, a new class of latency sensitive and compute intensive applications running across the network become possible. High-quality AR and VR video rendering requires a powerful computing platform. With Render Streaming, the complex rendering is done on a server which then streams a video to the client side. Battery driven, low-end devices can display streamed video using far less on-board power and compute resources. For VR and AR, positioning, movements, and other interactions on the client side need to be integrated in a tight control-loop with rendering to properly impact the displayed video.

The user experience is highly dependent on low latency in this loop. 5G and Edge Computing is a promising concept, enabling this type of latency-sensitive applications. The goal of this project is to demonstrate a Render Streaming application (AR or VR) running on the 5G network at Luleå University of Technology, using the RISE ICE Edge Computing platform for rendering. In addition to selecting a compelling application scenario and client devices, this also includes researching available toolchains for rendering and streaming, using Kubernetes on the ICE Edge Computing platform (with GPUs). One possible building block could be a computer vision framework developed by RISE.

The framework is based on Gstreamer and Nvidia Deepstream to leverage modern GPU hardware and hardware acceleration for media processing and machine learning inference. The framework also makes it possible to easily deploy computer vision applications on Kubernetes. Another goal of the project is to identify bottlenecks and optimize the performance. The thesis should investigate tradeoff between quality and performance, e.g. find bottlenecks and investigate maximum possible framerate and video resolution with respect to network latency and edge server performance. As the control-loop is heavily dependent on network latency, the thesis should particularly benchmark and investigate network latency implication on AR usability.


  1. Review relevant scientific publications.
  2. Review existing AR/VR open source frameworks that can be used as building block.
  3. Define a use case and refine research questions.
  4. Develop a working prototype that demonstrates the concept of AR streaming based on 5G edge computing.
  5. Define performance metrics and perform experiments to measure AR performance on the LTU 5G testbed.
  6. Make recommendations for further development.


Scope: 30hp, 1 semester full time

Start date January 2022 (flexible)

Location: Luleå

Compensation: A scholarship of 30,000 SEK (1,000 SEK per hp) is granted upon approval of the final report

This project will require skills for and an interest in video processing, prototype development, user experience, networking. As with any thesis project, a scientific approach will be essential.

RISE gather skills and personalities required for the transition to a sustainable economy, society and planet. We devise solutions that can make a difference here and now, while working on research areas and technologies that will be vital for tomorrow. Here, perspectives, expertise and people come together, from technology nerds to ICT experts and computer scientists to microbiologists. We are RISE and we have an offer that is hard to refuse. Challenging assignments that do good.

Read more at ri.se

For questions and further information regarding this project opportunity, reach out to recruiting manager Emil Svanberg, +46 10 228 46 28.

Applications will be submitted via We like students where you can also find other interesting opportunities.

Om jobbet




Visstidsanställning 3-6 månader

Job type

Student - examensarbete/praktik


Emil Svanberg
+46 10 228 46 28



Sista ansökningsdag


Skicka in din ansökan