Link Reliability-Based Adaptive Routing for Multilevel Vehicular Networks

In multilevel vehicular ad-hoc network (VANET) scenario, dynamic vehicles, complex node distribution and poor wireless channel environment deteriorate the reliability of routing protocols. However, for the key issues of relay selection, existing algorithms analyze the wireless link performance without considering the influence of dynamics and shadow fading on location from GPS, as well as channel condition and buffer queue, which would lead to inaccurate link characterization and maladaptive to network variation. In this paper, we establish a dynamic link reliability model to portray the link complexity of multilevel VANET scenario, and propose a link reliability-based adaptive routing algorithm (LRAR) to improve the transmission efficiency. Firstly, we propose a Kalman filter-based estimation approach to amend GPS original data for precise location of vehicles. Then, we define link reliability to quantify the wireless link performance, and establish a multilevel dynamic link model (MDLM) to evaluate it. Moreover, to accurately describe the complexity of wireless links, we integrate the corrected GPS data and characteristics of multilevel VANET including vehicle dynamics, distribution hierarchy and shadow fading into the modeling of link reliability. Considering the difference of link state among diverse vehicles, a maximum deviation algorithm is introduced to adaptively calculate the weight of each parameter in the modeling. Finally, we formulate the routing decision as a multi-attribute decision problem, and select the link with highest reliability as transmission path. Simulation results demonstrate that LRAR outperforms the existing routing algorithms in terms of average end-to-end delay and packet delivery ratio.