Network Coding Based Reliable Communication In Wireless Networks

Recently, wireless networks have gained more and more attentions due to the price reduction of wireless devices and the easy deployment. However, wireless links are vulnerable to severe channel fading, interference and physical damage in wireless networks. Furthermore, compared with wired networks, the network energy and resources are more limited. Therefore, how to protect network from failures, ensure the reliable communication and utilize network resources and energy efficiently becomes a hot problem recently. Thus, the thesis mainly focuses on the reliable communication in wireless networks.In the past decades, many network protection schemes are proposed to protect from failures to ensure reliable communication, such as the power control schemes, Automatic Repeat reQuest(ARQ), Forward Error Correction(FEC) and network coding based protection schemes. Network coding technique is applied efficiently to protect network transmission and reduce transmission times.Network coding is a novel network technique. In wireless networks, we can apply network coding to improve network resource utilization, redundant transmissions, save network energy consumption and network resources. Network coding is, as it were, bringing a deep change for wireless networks. In this thesis, we mainly apply network coding to protect from failures for reliable communication and conserve network resources. The main research issues and contributions are summarized as follows:1. we study network coding based1+N protection in many-to-one transmission mode in wireless mesh networks in previous works, which the receiver can recover the packets from N sources (many to one transmission mode) on the fly when there is a single link failure on the paths form N sources to the sink. But previous network coding based1+N protection scheme can just be used in a restrict topology, in which any k sources will be connected to k+1routers, and each router just use one edge-disjoint path to transmit, which will waste network resources. In fact, if there are enough edge-disjoint paths,1+N protection can be applied to protect all sources. In our work, we extend the sufficient and necessary condition to verify whether the network can provide1+N protection and study how to extend network coding based1+N protection scheme proposed previously used in wireless mesh networks with a general topologies to exploit network resources efficiently. An auxiliary graph is proposed to extend the1+N scheme used in restrict topologies to a general topology and increase the network throughput greatly.2. we study an N+k protection scheme to protect against multiple failures in wireless mesh networks with multiple flows between a source and destination pair. When there are multiple failing paths from the source to the destination, how to recover resource-efficiently is the main work. Previous works can handle multiple link failures with retransmission or protect against the worst failures which will waste network resources a lot. Firstly, we study a network coding based protection against multiple failures to keep reliable communication and conserve network resources. N+k protection scheme is provided, which can both satisfy network protection requirement and improve the network throughput greatly. Secondly, the decidability of our proposed deterministic network coding scheme of N+k protection scheme has been proved in this thesis. Finally, the simulation results verify that our scheme can improve the throughput greatly compared with other transmission schemes.3. Flooding is one of the most fundamental services in wireless sensor networks. It facilitates sensor nodes to propagate messages across the whole network. Thus, we study network coding based flooding in asynchronous wireless sensor networks. Each sensor can receive different units. If we code probably the number of transmission will reduce greatly. Network coding based flooding is proposed to reduce the number of transmissions in duty-cycle wireless sensor networks. An auxiliary graph is proposed to transform the minimum flooding cost problem to an equivalent problem, which is to find minimum flooding tree in the auxiliary graph. The efficiency of the proposed system is validated by the simulations.