Research And Improvement On Channel Assignment And Routing In Wireless Mesh Network

Wireless Mesh Network (WMN) with its advantages of low cost, ease of maintenance, self‐organization and wide coverage has been a research focus in wireless network. The multi‐hop characteristics of WMNs can expand network coverage and multi‐channel technology can improve network throughput. However, the problems are that multi‐hop may make routing more complexity which leads to an increase of End‐to‐End delay, and the multi‐channel technology brings additional interference which makes End‐to‐End Quality of Service (QoS) worse. Routing scheme and channel assignment techniques which are correlative and influenced with each other are considered to be solutions to these problems.Some inherent problems in WMNs are analyzed and some novel or improved algorithms are proposed in this paper. The paper mainly focuses on researching protocols and algorithms on wireless routing and channel assignment.In the process of channel assignment, a traffic and interference model is introduced and improved to estimate the performance of communication links. According to the estimated results, the priority levels of nodes and links are established. Based on these priority levels the greedy algorithm is introduced and improved. Then an algorithm called P‐EGCA is proposed to enhance the effective throughput and to resist the ripple effect which always appears in channel assignment.For routing parameter, in order to achieve the objectives of power‐aware and energy‐savings in WMNs, the default airtime‐cost metric is modified and a new power parameter is introduced when routing based on IEEE802.11s. The power‐aware model is established and the transmission power of Mesh nodes is optimized. The results in NS‐3Simulator indicate that the new routing parameter and algorithm based on power‐savings can really save the whole network energy.For routing algorithms, two tasks are done in this paper.1) The Location‐based Coordinated Wireless Mesh Protocol (L‐CWMP) is proposed. It uses the tree routing and expands the priori‐tree to find the destination node. Then according to the position of destination node, it selects different types of routing protocols to find route before sending data. If the source and destination nodes are in the same Mesh subnet, the tree‐based routing protocol will be selected preferentially. Otherwise the on‐demand routing will be chosen. The simulation results show the routing efficiency of L‐CWMP is better than the RM‐AODV (Radio‐Metric Ad hoc On‐Demand Distance Vector) which is an on‐demand protocol and OTR (Optimal Tree‐based Routing) which is of tree‐based routing.2) A new routing algorithm providing QoS guarantees is proposed. The End‐to‐End delay is taken into account as a constraint in routing protocol. The RREQ (Route Request) message in traditional AODV is adjusted to adapt the delay constraint. Then the current End‐to‐End delay and the maximum delay an application can tolerate are compared to determine whether to accept the current connection. The simulation results show that the proposed QoS‐AODV algorithm has a better performance in routing cost, End‐to‐End delay and packets loss rate compared with AODV‐MR.Channel assignment and routing are also combined to improve the performance of multicast in WMNs. With QoS delay constraint, the strategy of joint channel assignment and routing based on genetic algorithm (GA‐JCAR) is proposed. It is proved that this GA‐JCAR strategy can reduce interference and End‐to‐End delay in multicast networks.For meeting different requirements, the paper builds corresponding simulation environments for all of these improved or novel algorithms and protocols. Then the theoretical analyses are verified sufficiently by the simulation works in these environments. Some future works are suggested based on the shortcomings of these algorithms and protocols.