Vehicular data offloading under uncertain journey planners

Self-information broadcast messages are recurrently send out by vehicular terminals to establish their presence and hence improve situational awareness in road transportation. Importantly, such information snippets could prove invaluable to cloud-based intelligent transportation system services that aim to improve safety, traffic efficiency, and offer a plethora of infotainment services. However, due to the bulk nature of the aggregated load collected by vehicular peers, the delivery of this traffic to the infrastructure proves to be a challenging task. The latter offloading problem is considered in this work, in which each individual vehicular station decides on the offloading strategy that maintains proper utilization of the available infrastructure network. For this purpose, a graph theoretic approach is followed in which stations first estimate the potential of eventually being serviced by an available infrastructure node and then decide on the best offloading strategy that minimizes the load variance between the current and subsequent serving nodes. Performance analysis results illustrate that the proposed solution provides substantial load balancing gains and is computationally efficient to implement in practice for arbitrarily large network instances.