Gabriella Convertino

A new UPnP architecture for distributed video voice over IP

In this paper we present a novel approach to video telephony where we are able to exploit the benefits of the Universal Plug and Play (UPnP) protocol: self-configuration, discovery and control of available devices connected to the network. At the time being, UPnP does not deal with Video Voice over IP (VVoIP) applications. Integrating a signaling protocol like the Session Initiation Protocol (SIP) with UPnP allows interaction between the videophone and other UPnP compliant devices, like a TV set or a media server. This interaction, for example, makes it feasible to display the callers image on the television screen, thus allowing the user to relax on the sofa and naturally converse. Another interesting scenario is the one where the user sends to the conversation partner a set of still pictures or short video clips acquired from a UPnP media server already available in the home premises. An advantage of this distributed approach is its cost effectiveness, since it allows to reuse existing UPnP devices also in the context of video telephony. The integration between SIP and UPnP has been achieved introducing a new UPnP device, the VVoIP device, which is a SIP User Agent able to interoperate with standard UPnP MediaServers and Medi-aRenderers.

Wireless adaptive video streaming by real-time channel estimation and video transcoding

A system composed of an MPEG-2 video transcoder to change bitrate, frame rate and frame size and a cross layer controller gathering information from physical, MAC, driver, and RTCP layers, calculating instantaneous network throughput, to optimize real-time adaptive A/V streaming over IEEE 802.11 is presented.

Flex-WiFi: a mixed infrastructure and ad-hoc IEEE 802.11 network for data traffic in a home environment

This paper presents a novel architecture, called Flex-WiFi, specifically designed for IEEE 802.11 home networks. Objective of the new approach is to improve legacy network capacity in such home scenarios. Network capacity is increased by allowing setting up multiple contemporary IEEE 802.11 connections using a single WLAN card. All connections share the same data link and physical layers but use different channels; the MAC is responsible to handle the channel switching as well as to allocate/de-allocate the direct link sessions. In particular we aim implementing a solution where stations can be associated to an access point and can have ad-hoc connections on a different pysical chanel at the same. The idea is to maximize the overall network capacity by transferring some traffic load from the infrastructure to a direct link session. Results have been provided by implementing Flex-WiFi in the Madwifi open source driver of the Atheros chipsets.

Adaptive FEC for 802.11 burst losses reduction

One of the challenges that video transmission over wireless networks (such as IEEE 802.11) must face is the packet loss that can heavily decrease the received video quality. What makes the challenge more difficult is the bursty nature of wireless losses. Many video decoders can mask the effects of small amounts of random packet loss, however, most are unable to successfully hide the effects of burst losses (periods with high loss data rate). Forward error correction (FEC) is a technique extensively adopted to increase error resilience. In unicast transmission, the FEC redundancy is usually added adaptively on the basis of a loss pattern feedback. In many cases the loss pattern is represented through a very simple statistics: the average packer error rate (PER). This paper proposes a technique to adaptively introduce FEC redundancy that exploits a feedback scheme based on a 4-state Markov model to describe the loss pattern. The model allows us to obtain a description of the loss pattern in terms of burst and gap length and density. The feedback is based on the RTCP eXtended Report - IP Video Metrics Report Blocks that includes statistics on error patterns, such as the average PER, burst length and density. Our simulation results show that the proposed method smoothes out the burst losses and outperforms solutions based on the average PER.

MMC05-1: An Adaptive FEC Scheme to Reduce Bursty Losses in a 802.11 Network

One of the challenges that video transmission over wireless networks (such as IEEE 802.11) must face is the packet loss that can heavily decrease the received video quality. What makes the challenge more difficult is the bursty nature of wireless losses. Many video decoders can mask the effects of small amounts of random packet loss, however, most are unable to successfully hide the effects of burst losses (periods with high loss data rate). Forward error correction (FEC) is a technique extensively adopted to increase error resilience. In unicast transmission, the FEC redundancy is usually added adaptively on the basis of a loss pattern feedback. In many cases the loss pattern is represented through a very simple statistics: the average packet error rate (PER). This paper proposes a technique to adaptively introduce FEC redundancy that exploits a feedback scheme based on a 4-state Markov model to describe the loss pattern. The model allows us to obtain a description of the loss pattern in terms of burst and gap length and density. The feedback is based on the RTCP extended Report - IP Video Metrics Report Blocks that includes statistics on error patterns, such as the average PER, burst length and density. Our simulation results show that the proposed method smoothes out the burst losses and outperforms solutions based on the average PER.

UPnP Architecture for Distributed Video Voice over IP Applications (Demos)

In this paper we present a novel approach to video telephony where we are able to exploit the benefits of the Universal Plug and Play (UPnP) protocol: self-configuration, discovery and control of available devices connected to the network. At the time being, UPnP does not deal with Video Voice over IP (VVoIP) applications. Integrating a signaling protocol like the Session Initiation Protocol (SIP) with UPnP allows interaction between the videophone and other UPnP compliant devices, like a TV set or a media server. This interaction, for example, makes it feasible to display the callers image on the television screen, thus allowing the user to relax on the sofa and naturally converse. Another scenario is the one where the user sends to the conversation partner a set of still pictures or short video clips acquired from a UPnP media server already available in the home premises. An advantage of this distributed approach is its cost effectiveness, since it allows to reuse existing UPnP devices also in the context of video telephony. The integration between SIP and UPnP has been achieved introducing a new UPnP device, the VVoIP device, which is a SIP User Agent able to interoperate with standard UPnP MediaServers and MediaRenderers.