P-DACCA: A Probabilistic Direction-Aware Cooperative Collision Avoidance Scheme for VANETs

One of the major challenges in Vehicular Ad hoc Networks (VANETs) is to find a stable and robust Cooperative Collision Avoidance (CCA) scheme to address the rising death toll caused by road accidents every year. This work presents a Probabilistic-Direction-Aware Cooperative Collision Avoidance (P-DACCA) scheme that takes into account realistic bi-directional traffic, which makes this work unique in the VANETs' collision avoidance domain. The scheme starts with formation of dynamic clusters, which becomes challenging due to bi-directional heterogeneous traffic and intra-cluster and inter-cluster collision avoidance. For clustering, we modify the k-medoids algorithm by incorporating Hamming distance as an additional metric for direction-awareness. After clustering, relative distances and speeds of nodes with respect to their expected states are computed. The scheme then estimates a collision probability on the basis of a node's expected state and provides an early warning when the probability exceeds a predefined threshold. For implementing a preventive measure, we introduce an adaptive Benign factor that computes the safe speed for a target node. The safe speed is encapsulated along with the collision probability into an early warning message for dissemination to the target node to avoid an upcoming threat. Simulation results demonstrate significant improvement of the proposed scheme in terms of cluster stability, reduced number of collisions, low latency and low communication overhead. (C) 2019 Elsevier B.V. All rights reserved.