A framework for Broker placement in vehicular ad hoc Networks

Vehicular networks are a classic example of mobile, peer-to-peer networks. The domain of intelligent transportation engineering (ITS) envisions a wide range of services over these vehicular networks. Supporting the different quality of service properties of these services requires effective dynamic resource management solutions in vehicular networks, which cannot be realized using vehicles alone due to factors such as varying vehicular mobility patterns, varying vehicular traffic density, and obstacles. Road side units (RSUs), which serve as event and data brokers among the moving vehicles, can alleviate this problem. To maximize the value proposition offered by RSUs requires their effective placement that holistically accounts for physical factors, such as traffic patterns and mobility, with cyber factors, such as communication protocols and messaging. To address these requirements, this paper presents a design methodology based on the principles of surrogate modeling wherein small-scale, micro-simulations of the ITS cyber-physical system are used to develop training points, which in turn are used to train a surrogate model. The surrogate model is subsequently used to make planning decisions for ITS.