Cost-effective vehicular network planning with cache-enabled green roadside units

Vehicular communication networks expect to accommodate the ever-increasing on-road wireless traffic by deploying roadside units (RSUs) and meanwhile exploiting existing wireless infrastructures. To achieve flexible deployment, energy-saving operation and low-latency services, a new type of RSUs, namely cache-enabled green RSUs are introduced, which can store popular contents locally and harvest renewable energy as power source. In this paper, we investigate cost-effective planning of heterogeneous vehicular networks consisting of conventional macro base stations and cache-enabled green RSUs. Specifically, the RSU density, cache size, and energy harvesting rate are jointly optimized to minimize network deployment cost, under the constraints of quality of service (QoS) requirements and limited backhaul capacities. For QoS guarantee, the lower bound of average data rate is derived in closed form by applying the theory of stochastic geometry, based on which the cost-effective network deployment scheme is proposed. Analytical results reveal the tradeoff between cache size and backhaul capacity, indicate renewable energy harvesting rate should be sufficient to support rush-hour demands, and also provide the optimal RSU density for the given vehicular traffic demands. Extensive simulations are conducted for validation. In addition, numerical results of optimal network planning are provided in details to offer insights into practical system design.