Loading...
Please wait, while we are loading the content...
Similar Documents
Source Coding for a Noiseless Broadcast Channel with Partial Receiver Side Information
| Content Provider | Semantic Scholar |
|---|---|
| Author | Timo, Roy C. Grant, A. J. Hanlen, Leif |
| Copyright Year | 2007 |
| Abstract | A transmitter communicates the outputsX, Y from a finite discrete memoryless source to two receivers via a noiseless broadcast channel. Each message is required at only one receiver, the receivers cannot cooperate, and one receiver has side information U . We show the achievable communication rates for this channel areR ≥ H(Y )+H(X|Y, U). Achievability is proved via a rate-split version of random binning. Cut-set outer bounds are not tight, so an alternate converse is presented. This problem is an example of a broadcast network, withm messages andm independent receivers each having different sideinformation. This broadcast network is motivated by a control information problem from mobile networking, and it generalizes the Wyner-Ziv, Heegard-Berger, and Sgarro problems. I. I NTRODUCTION Information theoretic analysis of a few simple networks has given tremendous insight into the structure of efficient communication strategies for large networks. The application of Slepian-Wolf distributed compression [1] in wireless sensor networks [2] is a good example. We present and solve a new distributed source coding problem that is motivated by the need to efficiently distribute network layer control information in mobile networks. Consider a network with a source that wishes to send multiple packets to multiple sinks. Typically, somerouting informationmust be included to ensure that each packet reaches its intended destination. For example, in the Internet Protocol (IP), packet headers contain the destination IP address. Although such route information assists with the network layer operation, it is an overhead that reduces the capacity of the network for payload traffic. It is of interest to develop efficient compression strategies for route information, and to determine the corresponding information theoretic limits. Mobile ad-hoc networks in particular may require significantly more network layer signaling, and information theoretic limits on routing information may be of interest in determining the impact of mobility on network capacity. We are interested in scenarios like the one just described, where the different routing information required by each relay (router) must be generated by the source, and encoded into a common message provided to all relays connecting the source and sink. Each successive relay must be able to independently make the correct routing decision (e.g. where to forward the packet), based on this common message. Furthermore, R. Timo and L. Hanlen are with National ICT Australia, and The Australian National University. Email Roy.Timo@anu.edu.au, Leif.Hanlen@nicta.com.au. A. Grant is with the Institute for Telecommunications Research, the University of South Australia. Email Alex.Grant@unisa.edu.au. This work was funded by a joint ARC Communications Research Network/National ICT Australia travel grant. National ICT Australia is funded through the Australian Government's Backing Australia's Ability initiative, in part through the Australian Research Council. each relay possesses different side information relevant to its routing decision. For example, this side information may consist of routing tables, knowledge of the local network topology, and knowledge of the hierarchical structure of IP addresses. Side information is the key ingredient in the source coding problems considered in this paper. We are interested in the degree to which side information reduces the amount of routing information required. Our simplified information theoretic abstraction of this route information problem is a noiseless broadcast channel with receiver side information, Figure 1. The common transmitted message represents the routing information in the packet header. We consider a noiseless channel, which is a reasonable assumption for network layer communication. Each of the N receivers corresponds to a relay, having its own side information. Note that our model does not allow for modification of the packet header by relays (respecting endto-end connectivity at the network layer). Whereas SlepianWolf [1] consider distributed encoding of multiple sources and a single decoder, we consider a kind of dual – joint encoding of multiple sources, and distributed decoding. Source Encoder Decoder 1 |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://users.cecs.anu.edu.au/~leifh/ftp/ausctw07/pdf/ausctw2007_part16_86_90.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
