Delay-tolerant routing and message scheduling for CR-VANETs

Abstract

Cognitive radio vehicular ad hoc networks (CR-VANETs) can solve the problem of the limited spectrum resource and improve the vehicle-to-vehicle (V2V) communication efficiency. Nevertheless, the fast moving characteristic of vehicles and the occasional sparse density of vehicles lead to the intermittent CR-VANETs. Recent literature of routing in CR-VANETs focuses on a fully interconnected environment yet fails to address solutions for non-real-time applications. To fill this gap, we propose the delay-tolerant routing and message scheduling (DTRM) schemes for non-real-time applications, aiming to maximize the delivery ratio and decrease the delivery overhead in CR-VANETs. First, we propose a routing scheme that builds a concurrent forwarding set and designs a forwarding strategy. In this set, we select relay candidates by considering both the CR channel availability and the V2V contact duration. Moreover, we evaluate the priority of the relay candidate to provide a reliable forwarding strategy. Second, based on the routing scheme, we propose a message scheduling scheme that combines an optimized-binary-tree replication algorithm and a buffer management policy. The proposed algorithm can spread messages and terminate the replication process quickly. The buffer management policy facilitates the tradeoff between the delivery ratio and the delivery overhead, taking into consideration the node priority, the message tag, and the message remaining time-to-live. Simulation results show that DTRM presents a higher packet delivery ratio and a lower overhead ratio than the counterparts. It presents on average gains of 25.1% and 10.9% in terms of delivery ratio and overhead ratio, respectively, when compared with four CR-VANET routing schemes. It also presents on average gains of 35.6% and 34.1% in terms of delivery ratio and overhead ratio, when compared with three message scheduling schemes.