| EPSRC Reference: |
EP/C538692/1 |
| Title: |
Fuzzy-Logic Transcoded Video Stream Controller |
| Principal Investigator: |
Ghanbari, Professor M |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Computing and Electronic Systems1 |
| Organisation: |
University of Essex |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 May 2005 |
Ends: |
30 September 2008 |
Value (£): |
228,861
|
| EPSRC Research Topic Classifications: |
| Artificial Intelligence |
Digital Signal Processing |
| Image & Vision Computing |
|
|
| EPSRC Industrial Sector Classifications: |
| Communications |
Creative Industries |
| Information Technologies |
|
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
The Internet's growth has encouraged new multimedia applications, such as DVD-video streaming, delivery of sports and news clips, and the exchange of personal video clips (peer-to-peer streaming). The Internet is a packet network in which data streams are split into packets. As compressed video is passed across a fixed network, it will compete with other packets for use of links. The most obvious competitors are Web page packets. However, increasingly MP3 files are being transferred between users, and, of course, so are other video streams. If a link is occupied then a packet is stored in a router's queue for later transfer. If the buffer holding that packet becomes full then packets will be dropped. The effect of delay is that arriving video frames cannot be decoded at regular intervals so that the viewer will see a juddering display. Buffering can reduce the impact at the receiving device, but on smart phones the buffer memory is limited and anyway a large buffer might cause a perceptible delay before a clip starts up. More serious is excessive loss of packets, because this may prevent video frames being displayed, resulting in a significant loss of quality, especially in a sports clip with rapid motion.To avoid this effect of packet delay and loss (network congestion), the process of transcoding is applied to alter the rate at which video is transferred, reducing the number of packets according to the congestion level. The problem of how to decide upon the transcoding output rate remains, which is the essential issue that the research in this proposal concerns itself with. The existing protocol for the transfer of packets (TCP) provides reliable delivery but trades reliability against (often unacceptable for video) delay. A variety of congestion-control methods have been suggested to allow delay to be traded-off against reduced video quality. However, though these methods were only proposed in the mid-90s, they may already be 'elderly', because Internet expansion is such that the character of the constituent networks is already changing. Some changes are: the arrival of error-prone wireless networks; mobile devices with limited playback-buffer space; and the trend towards peer-to-peer file transfer. The main concern of the designers of these congestion-control methods was to avoid upsetting TCP traffic by acquiring more of a network's capacity as TCP reduces its transfer rate. Mathematical models were devised that closely mimicked TCP's behaviour (though allowing packet loss), as essentially Internet protocols behave in a cooperative manner.Existing congestion control methods may be unsuitable for the emerging mix of network and traffic types. Our experiments have suggested that a nonmathematical method of control may even be more efficient in delivering higher-quality video, without having an adverse effect on existing traffic. This method is based on fuzzy logic, which is a form of artificial intelligence. Fuzzy logic may also be robust to future changes in the way the Internet is formed. Given that the feedback channel that governs the response of the controller uses the same path as the outward video stream, then the controller can, as it were, upset its own response (it is non-linear). This implies that as differing network characteristics occur a mathematical model will become difficult to design. For example, some existing methods rely for their feedback on knowing the number of packets that have been lost, yet loss is an unreliable indicator of congestion on a wireless network, as packets may be dropped due to adverse transmission conditions. As fuzzy logic is most appropriate when there are non-linearities, such adaptive methods may prove to be superior. To move nearer to a robust multimedia service a fuzzy transcoder control unit should be implemented. Recently available large-scale FPGA devices, which are a type of modular micro-chip, can be configured at will, making them ideal for prototyping.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.sx.ac.uk |