Research On Channel Assignment Mechanism Of Vehicular Network Based On RoF-DAS Architecture

In recent years, the intelligent transportation system(ITS) has attracted wide attention in many countries for its reduction of traffic load, guaranteed traffic safety and facilitation of traffic flow. As a promising Intelligent Transportation System technology, the vehicular network provides the vehicles with a variety of safety-related applications(i.e., accident warning and road traffic management) and other applications(i.e., Internet access and multimedia data transmission). Due to the uncertainty of vehicular environment, the fast-moving vehicles face a challenge of communicating with other vehicles and the roadside infrastructures efficiently. Additionally, due to the large volume of traffic, the increasing interference and contention among vehicles seriously degrades the performance of the vehicular network. In order to satisfy the demand for high quality transmission and mitigate such interference, an optimal radio access and efficient frequency resource management is strongly required.To solve those problems, we propose a new vehicular network that integrates Radio over Fiber(Ro F) technology with Distributed Antenna System(DAS) for future ubiquitous communication and increasing demand of service bandwidth. Based on the Ro F-DAS vehicular network, a Centralized Cluster-based Muti-channel MAC(CCM) scheme is proposed for wireless access in vehicular environment which integrates the clustering with contention-free/-based MAC protocols. Firstly, clustering scheme is proposed based on fast affinity propagation algorithm according to vehicle similarity degree and vehicle node degree. The cluster head vehicle is equipped with two sets of transceivers which can operate simultaneously on inter-cluster channel(ICC) and cluster-range channel(CRC) respectively. Secondly, to reduce the disorderly competition and interference in the cluster, the cluster member vehicles transmit safety-related and unsafety-related messages based on polling access scheme and time division multiple access scheme respectively. Thirdly, the chosen cluster head vehicles communicate with center control stations(CCS) or other cluster head vehicles using competition-based method which is flexible and less sensitive to the high mobility.On the above basis we design an innovative RF assignment mechanism which includes ICC and CRC allocation mechanisms. Firstly, considering the frequent shift betweendifferent RAUs for the fast-moving vehicles, We integrate several RAUs in an RAU group connecting to an RFSP in order to reduce the frequency of switching between different RFSPs. The size of the RAU group can be adjusted adaptively by CCS according to the vehicle density and average speed. Secondly, the CRC allocation scheme is proposed based on the instant interference of the cluster and the load information of the channel. By exchanging information using beacon frames, cluster-head vehicle constructs a localized graph of neighbor clusters and chooses the channel with minimized interference. Thirdly, the ICC allocation is put forward based on site-specific cognition of vehicles to capture the vehicles’ utility function within the cluster timely, thereby improving the RF utilization of the entire network and load balancing globally. Lastly, we compare the performances of the proposed channel allocation scheme in Ro F-DAS vehicular network and the traditional distributed channel allocation method based on measurement in VANET using OPNET Modeler. Numerical simulation results demonstrate that the proposed channel allocation scheme outperforms the traditional distributed channel allocation method in terms of normalized throughput of total network, delay and packet drop rate, which is especially obvious in the hotspot areas.