Research On Partially Overlapped Channel Assignment For Wireless Mesh Networks

The rapid development of communication technology has changed our lives. As a wayof solving “last mile” bottleneck problem of Internet, wireless mesh networks are underincreasing concern. Wireless mesh networks have the advantages of high capacity, highspeed, low cost and robustness, which can be widely used in home broadband network,enterprise network and transportation system. Therefore, the study on wireless mesh networkperformance has become a hot topic in wireless mobile communication area.Interference is the main factor leading to network performance degradation in wirelessmesh networks, and multi-radio multi-channel technology is the most simple and effectivemethod in reducing network interference, so it’s particularly important to design reasonableand effective channel assignment algorithms to improve the performance of multi-radiomulti-channel wireless mesh networks. Existing channel assignment algorithms for wirelessmesh networks are mainly focus on orthogonal channel assignment, however, when thetraffic become very dense, it’s difficult to avoid assigning the same channel to adjacent links,the co-channel interference will decrease network capacity. With the development ofsoftware defined radio and cognitive radio, the interference between partially overlappedchannels has already been effectively controlled, so the network performance can beimproved by carefully designed partially overlapped channel assignment, therefore, theresearch of how to use partially overlapped channels for data transmission is of greatsignificance.Existing traffic-aware unicast partially overlapped channel assignment algorithms ofunicast are not suitable for the network with dense traffic, in this paper, a new traffic-awarepartially overlapped channel assignment (TA-PCA) is proposed. Link mutual interferencematrixes are proposed to model interference between different links with respect to differentchannels, based on which, multiple rounds of interference-free channel assignment areperformed for links that need to be assigned channels, then static link scheduling is used toguarantee network connectivity. After that, TA-PCA optimizes link scheduling to furtherimprove the network performance.A source node can transmit information efficiently to multiple destination nodes byusing multicast communication technology, compared with unicast communication,multicast communication can save channel resources effectively. Aiming at working out theproblem that existing multicast partially overlapped channel assignment may suffer fromhidden channel and randomly selects one channel from the channels may not give the bestperformance, a new partially overlapped channel assignment for multicast (M-POCA) isproposed. The algorithm takes link set as channel assignment unit, interference-free channelassignment are performed for link sets under full consideration of mutual interferencebetween link sets. After that, the algorithm performs complementary channel assignment forthe link sets which are not assigned channels with the goal of minimizing the overall network interference, in order to enhance the network connectivity.The unicast and multicast partially overlapped channel assignment are simulated onnetwork simulator version3. Simulation results showed that TA-PCA and M-POCA canachieve good performance in terms of the network throughput, end-to-end delay and packetloss ratio.