A cross-layer design for next generation wireless networks: V2V perspective

Recent advances in the automotive industry enabled us to build fast, reliable, and comfortable vehicles with various safety features. Roads are designed and made safer than ever before as well. Nevertheless, analyses and reports show that traffic accidents still remain one of the major causes of death and/or serious injuries around the globe. are expected to minimize the total number of accidents, if not prevent them completely. Safety is not the sole objective of . Routing optimization, green environments with reduced fuel consumption and carbon emission, and infotainment applications are also prominent targets of . It is clear that all of these objectives, envisions, and services require some sort of a communications network structure. Note that two key components of any transportation system are vehicles and the transportation infrastructure. Therefore, it is easier to analyze the network structure required by in terms of and networks. A network in represents a set of physically close–by vehicles which are communicating with each other within a local geographical region. network, on the other hand, consists of vehicles that communicate with the transportation infrastructure. Due to the high–level mobility, signaling in both and networks is established via wireless links. Generally, is considered to be supported by a backhaul over the network so that it is connected to the core or backbone network. Among and networks, networks receive slightly more attention compared to networks due to the following reasons: First of all, network topology for networks is dynamic and of transient nature because of high–level of mobility. This implies that network entry, establishing connection, and maintaining a high–level of are relatively difficult tasks as compared to those in traditional terrestrial communications networks such as cellular networks. Second, network traffic consists of several types of messages including emergency–related data with various requirements. From this perspective, information flow, data integrity, authorization, and security become life–critical concerns. In addition, in case there are multiple networks in the vicinity, relaying especially the critical emergency–related messages from one node to another needs to be considered very carefully. In this regard, establishing a connection between transient networks along with the aforementioned information flow, data integrity, and security concerns is a serious challenge for the networks. Note that none of the concerns listed here poses severe problems for the networks given that the transportation infrastructure can handle the signaling seamlessly. In order to tackle the problems and concerns listed above, generally the traditional layered architecture is adopted in the literature for networks. Although it is a very powerful and successful strategy, layered architecture falls short in solving some critical issues especially in networks such as dynamic resource allocation. This points out that a cross–layer approach could provide different perspectives while benefiting from the layered architecture. At the end, it should be stated here that the standardization efforts are not mature enough yet for networks. For instance, network entry procedures and the non-emergency/service channel selection mechanisms are not explicitly defined in the draft version of the standard. This automatically indicates that there are some design gaps which could be supported by the findings of research carried out in this field especially with the cross–layer support. Hence, in this dissertation, a novel cross–layer predictive channel selection mechanism is proposed for networks in order to minimize the average number of collisions. Both and layers are incorporated into the cross–layer design. At layer, first a novel, fractional rate sensing mechanism is proposed, which reduces the total number of computations in order to obtain sufficient statistics for the decision. Next, the necessary condition for the optimum predictive sensing strategy is derived and validated by the empirical data obtained by field measurements. It is also shown that any linear predictive strategy outperforms the general Markovian–based prediction schemes under various traffic load scenarios in case the derived necessary condition is satisfied. Finally, a protocol which is developed based on master–slave architecture along with a nomination procedure operating on a single universal broadcast channel is proposed at the layer. The proposed protocol is fed with the output of the layer predictive channel selection mechanism and completes the network entry procedure for networks.