Wireless Vehicular Check-in Protocol at Urban Road Intersection

Abstract

Intelligent organization of the traffic flow at an urban intersection is fundamentally important in intelligent transportation systems. Because of their low-cost deployment and convenient maintenance, wireless devices attract more attention against traditional devices, e.g., ground loops and road cameras. This paper targets to provide a vehicle-efficient check-in at a roadside unit (RSU) installed at an urban intersection through wireless communication, which means that each vehicle exchanges information with the RSU while traveling past the intersection. While this seems simple, it is nontrivial because fast-moving vehicles limit the lifetime of their packets and the RSU suffers from mutual interference caused by other mobile devices. Especially, the communication background environment is complex and highly dynamic. To address the challenges, we design a wireless check-in protocol with flexible parameters on the basis of the IEEE 802.11 protocols. The protocol leverages a vector auto-regressive and moving average model and reinforcement learning to dynamic environment adaptation. Moreover, a fine-grained tuning method based on an exponential smoothing algorithm is applied on the RSU side to further improve the check-in efficiency. We evaluate the proposed check-in protocol by combining the real urban road intersection data, collected in Shenzhen and Shanghai, with a simulation of urban mobility (SUMO) and an NS-3 network simulator. Its usage brings about 10% improvement in both check-in ratio and packet reception ratio in comparison with the IEEE 802.11p. It also can act as a basic component for various real-time information services.