VS:products

Our Potential Products. Back to Main Page

We need to provide short, medium, and long term products. For each product, we need to estimate its complexity (time to market), potential customers, competing products, our edges, links to our software design, hardware, etc.

Short Term 0.5 - 2 Years

Product S1: Micro Mobile TV Base Station for Unlicensed Use

Our first product is a low-cost workstation that allows small-scale deployment using unlicensed spectrum. Potential customers include universities, companies, communities, museums, and zoos, which want to efficiently (in terms of costs) deliver video streams to users within small geographical areas. For example, museums and zoos can provide visually-enhanced guided tours using our base stations without huge capital investments such as licensing a spectrum. In addition, customers may receive the programs using their own cellular phones.

This product is similar to our MM08 mobile TV testbed. Additional enhancements include: (i) mechanisms to query white space database, (ii) hardware to sense co-channel interference (need to check the FCC standards), and (iii) integration with some mobile TV capable devices, i.e., implementing a client program. Using a 5MHz channel within a short broadcast range, our base station can transmit 40 QVGA programs, where each program has a coding rate of 512 kbps. Concurrently reading and encapsulating 40 programs should be feasible to recent quad-core CPUs and multi-Gigabit disks.

We are not aware of any base stations that supports broadcasting over white space channels yet.

Product S2: VideoCast User Generated Content over Mobile TV Networks

While YouTube has been very successful on PC users, it is less popular on mobile devices for two reasons. First, mobile devices have low-speed wireless links. For example, existing HSDPA networks have a theoretical maximum bandwidth of 14 Mbps, which is shared among all users in the same cell. Second, mobile devices have small screens, which make browsing and searching for clips difficult. Effectively searching videos is important for user generated databases like YouTube due to the large number (millions) of videos in them.

We design a VideoCast platform to address the above two issues, and broadcast the most demanded videos to the most users. This platform consists of a VideoCast server that schedule the videos for broadcasting, and a VideoCast client running on mobile devices. More precisely, the client collects user preference from individual users. The user preference may contains regular expression patterns of video names and description, videos uploaded by certain users, some video channels, and even video tags. To support a large number of users, the client reports user preferences infrequently, in the order of hours. In addition to preference, each client may inform the server its status such as reception quality, battery level, and storage space. Client status messages are short, and thus can be sent more frequently. Nevertheless, the client should implement some throttling algorithm to avoid overloading the VideoCast server.

The VideoCast server uses the user preference and client status to compute the broadcast schedule in order to maximize the user experience. At a high-level, a video clips that appears at the top of the user preference lists for more active users, it should be scheduled earlier than less demanded videos. Since each video is sent to many mobile users, the VideoCast platform is bandwidth-efficient. Moreover, the VideoCast server will send the most demanded videos based on a user preference configured off-line, which saves users a lot of time on browsing and selecting videos using small screens of mobile devices.