Gelecek nesil araçlar arası haberleşme ağlarında bağlantı ve güvenilirlik modellemesi

Due to the vehicular network technology which enables to obtain online and accurate information between vehicle to vehicle and vehicle to infrastructure, the amount of vehicular applications has been increased. To support many vehicular applications, the U.S. Federal Communications Commission has allocated 75 MHz in the 5.9 GHz frequency band for Dedicated Short Range Communication (DSRC). However, due to the increasing use of vehicular applications and high traffic density, specifically in urban environments, DSRC channels have not satisfied to growing bandwidth demands by causing a crucial problem; spectrum scarcity. This problem has been addressed by a new communication paradigm, called Dynamic Spectrum Access. In this thesis, at first, we concentrate on the connectivity problem in Vehicular Dynamic Spectrum Access Networks (VDSANs). Our objective is to determine best available channels to maintain continuous and robust network connectivity. In this thesis, we propose a novel queuing theory based framework for VDSANs. Specifically, our proposed framework uses queuing theory analytics to model the dynamic behaviors of vehicles and obtain high satisfaction ratio of vehicles. Moreover, we propose six novel dynamic channel selection algorithms to provide minimal channel switching ratio while conserving the network connectivity. In addition to connectivity analysis and modeling in VDSANs, we also analyze the reliability of broadcast messages in vehicular networks. In IEEE 802.11p based broadcast vehicular communications, there are no traditional request-to-send (RTS)/clear-to-send (CTS) handshaking and acknowledgment mechanism to guarantee a reliable communication. Hence, hidden terminal problem, transmission collisions and channel fading arise as main communication challenges in vehicular networks and all these challenges dramatically reduce the reliability of broadcast vehicular communication. Therefore, in this thesis, we also analyze the reliability of broadcast vehicular communication by proposing a novel power calibration model implemented into RSUs. More specifically, the proposed model is used to estimate the reliable and effective transmission power of hidden vehicles with Kriging spatial interpolation method. Kriging enables to find an optimal spatial prediction which estimates the unknown values from the observed data. Kriging interpolation method is used to predict the right amount of transmission power of hidden vehicles in vehicular networks so that we can calculate the hidden terminal radius depending on traffic density to prevent hidden terminal problem.