Video streaming is fast becoming the dominant source of data traffic in wireless networks. In order to avoid overburdening the cellular infrastructure, this project investigates new systems that exploit caching to provide high throughput with minimal infrastructure cost. This is pursued by exploiting the asynchronous content reuse of on-demand video, i.e., that most people watch the same videos, though at different times. Two approaches to such caching at the wireless edge ar considered, either deploying helper stations (femto-base stations with a cache, but no backhaul), or caching on user devices together with highly spectrally efficient device-to-device (D2D) communications. The investigations will consider the following main topics: (i) develop policies for optimizing video quality in wireless streaming, (ii) investigate scheduling for single-hop transmission between neighboring devices, (iii) investigate coding for D2D transmission for combining D2D with coded multicast schemes. To complement these theoretical investigations, realistic operating conditions (channels, user distributions, etc.) will be analyzed through simulations.
The result of this work will allow more spectrally efficient transmission of wireless video. This will increase overall video consumption, which is beneficial to the movie/entertainment industry, a key US industrial sector. Furthermore, the suggested systems can also free up many wireless resources that can be more efficiently used for other purposes.
