Research to Empower and Optimize Tomorrow's Interactive Services Delivered over the Internet
Supervisor: Associate professor Niklas Carlsson, ADIT, Department of Computer and Information Science
Context and Research Area:This project aims to provide novel, adaptive, and efficient solutions for delivering tomorrow's exciting and highly interactive services over the internet. Services of particular interest include interactive video streaming (e.g., branched videos such as "Black Mirror Bandersnatch", in which the users select their own path through a video, 360-degree video in which the users can look in any direction, free-view-point streaming in which the users control the position of the camera and can select from what angle they want to view a previously recorded scene), virtual reality, augmented reality, and mixed reality.
In addition to help satisfy users' increasing expectation of personalized and interactive service, these services have the potential to disrupt how we do business and interact with people and our environment. However, to reach their full potential, such services need to be effectively provided to end-users whenever and wherever they are, and ideally when using any end-system device. This goal presents major challenges. For example, today, many of the above mentioned services require significant computing resources available on the end-user devices themselves and the services often do not work well over the internet. To enable the next generation of interactive services, new solutions are needed to address a number of fundamental problems common for these types of services, including how to best hide heterogeneous and time varying bandwidths, delays, and other resource constraints from the end users. Due to the human component, many of these challenges must also be addressed both from a technical and from a user perspective, taking into account the users' quality of experience (QoE).
To address these problems challenges, in this project, we will use a range of methodologies to (i) better understand and improve existing solutions, and to (ii) design, implement and evaluate solutions for future services. For the biggest impact, we will consider solutions from multiple perspective, including from an end-user perspective, a network provide perspective, and a content provider perspective. For each of these contexts, frameworks and optimized solutions will aim to provide the best possible service, hide network delays, mask resource constraints, and optimize clients' performance and perceived QoE, while taking into account the heterogeneous network conditions and device capabilities of each client, the interests and preferences of each user, and the high delivery costs associated with these services. While challenging problems in themselves, if addressed carefully, we believe that our frameworks and solutions have the potential to enable fundamentally new services and business models that go well beyond what is possible with existing technology.
As a PhD student, you will work with myself (Niklas Carlsson) and some collaborator partners. While the research would be adapted based on the skills of the candidate, the research is likely to involve a subset of the following methodology approaches:
- the design, development, and implementation of practical solutions within open-source frameworks,
- experimental evaluation of both existing and improved solutions,
- user studies based on instrumentation of our solutions,
- large-scale measurements of public services/systems using public APIs or custom-made measurement tools,
- data-driven analytics, applied machine learning and/or the use of other statistical methods, and
- develop mathematical evaluation frameworks, optimized protocols, and derive fundamental performance/cost bounds.
We are looking for a highly motivated candidate with a genuine interest in learning, discovering, and carefully exploring new things. As the ideal candidate, you are expected to have the ambition of doing high-quality research and to strive towards making impactful contributions. This typically includes a desire to understand and improve actual systems; e.g., by implementing new solutions into existing systems, collecting and working with (often) large-scale datasets from the real world, or performing comprehensive instrumentation and experimentation to demonstrate the value of different solutions. Other desirable properties are that you have strong programming skills, problem solving skills, and English writing skills. As you likely are to work with at least 2-4 of the 6 example methodologies mentioned in the above paragraph, it is desired that you have some experience working with at least some of these different methodologies. In your application, it would therefore be great if you highlight what you consider to be your primary skills and interests. Here, you can also reflect on what methodologies you (at this time) think would be your preferred choice and after some training may allow you to achieve most success. Finally, to work on this project it is also considered an advantage (but not a requirement) that you have some background (e.g., taken courses or have other relevant experience) with at least one of the topics multimedia technologies, game development, measurements, data analytics, HCI, networking, web technologies, or have strong mathematical skills.