Unmanned aerial vehicle (UAV) has aided in many communication networks such as cellular and device to device network. Use of UAV in the area of vehicular ad hoc networks (VANETs) is an area which has substantial potential. We present such a framework in which UAV provides assistance to the road side unit (RSU) for serving the vehicles. The framework uses carrier sense multiple access scheme (CSMA) with capture effect. This access mechanism is well famous for performance degradation due to its contention-based random nature. We model the behavior of the vehicle by using two dimensional Markov chain; evaluating the state as well as transition probabilities. Parameters such as signal to interference plus noise ratio (SINR) threshold and the access probability are optimized. Access probability is the probability of a vehicle using certain number out of total number of available channels. We present the RSUUAV framework to enhance the overall throughput and number of successful packets when compared with conventional vehicle-to-infrastructure system. We study the effect of different path loss component of ground-to-air link and ground-to-ground link of UAV and RSU servers respectively on the overall throughput. The transmission power variation is carried out with a combination of relative distance criteria and its scaling factor. Scaling factor is used for compensating the additional height of the UAV when compared with height of the RSU. We numerically find the best possible combination and scaling factor parameter. The proposed framework offers throughput improvement of 37.5%, 24.4% and 8.97% for vehicular density of 60, 120 and 180 respectively when compared with the RSU-only framework. Whereas, the proposed power scheme offers a throughput of upto 20.49% when compared with the constant power case.