Heterogeneous Architecture for DSRC/LTE Vehicular Communication Networks Based on the ITS Reference Architecture with Fuzzy Logic for Decision-Making

Abstract

Methods to integrate different access technologies such as heterogeneous communication architecture remains a challenge. Among a variety of V2X-supporting access technologies, Dedicated Short Range Communication (DSRC) and cellular communications, e.g., Long Term Evolution (LTE) is promising reliable and efficient vehicular communications. DSRC has been designed to allow for direct low-latency communications among different vehicles (i.e., V2V) and between vehicles and roadside units (RSU). On the other hand, DSRC suffers from link quality degradation with the presence of buildings and vehicles, especially in urban areas, where channel collisions become serious when vehicle density is high. LTE networks can provide wide area coverage and are favorable to bandwidth-greedy applications, which require high data rates and reliability. Considering the relatively high end-to-end delay for message transmission due to the long transmission time interval (TTI) current LTE networks have drawbacks regarding latency to support highfrequency safety-related information exchange among vehicles in local areas. Combining LTE and DSRC approaches as a heterogeneous solution is essential to fast introduce V2X services for future automated driving. By this intelligent Vertical Handover (VHO) algorithms are needed to ensure seamless connectivity of vehicles to the best network at a particular point in time. For this paper, we propose a 3- input fuzzy-logic-based VHO scheme for Heterogeneous DSRC/LTE Vehicular Communication Networks. The simulation results in this paper show that the proposed 3- input fuzzy-logic model can be able to make decisions and select the best network to connect to base on the network with the highest HF value. The 3-input fuzzy-logic model was also tested in a vehicular highway scenario simulation. The proposed model shows optimized communication results in terms of the mean end-to-end delay against that of the literature.