In recent years, due to the development of various applications that can be consumed within connected vehicles, the amount of vehicle network traffic in Vehicle Ad-Hoc Networks (VANET) has been increasing enormously. And also, in VANET, the network reconnection occurs frequently due to the high-speed movement of vehicles and frequent changes in the direction of movement. For this reason, the service in VANET is hard to serve with reliability and high data rate.
Therefore, in this paper, we propose a pre-caching scheme in which the data requested by the user driving the vehicle while on the move are brought to the expected travel path of the vehicle in advance.
To predict the direction of vehicle movement, we designed the pre-caching scheme, which is based on the usage frequency of Road Side Units (RSU) and the direction of vehicle movement.
The reason why we use the usage frequency of the RSU is that it takes into account a specific pattern in which the vehicle moves according to the characteristics of the road.
For example, the usage frequency of RSUs to which the vehicle is connected for network use is counted according to the direction of movement of the vehicle so that the frequency of RSUs with a high frequency among adjacent RSUs has a high possibility of movement.
Besides, the vehicle''s direction of movement is derived by calculating the slope of the vehicle''s direction of movement through the coordinate values of the RSU previously used by the vehicle to derive a direction with a high possibility of movement of the vehicle. The simulation was designed and experimented with by implementing a vehicle movement model through actual New York City taxi data. Even if data such as GPS, speed, or travel habits were not considered from the vehicle, it was confirmed that the vehicle''s direction of movement could be predicted only by RSU on the road. The vehicle''s caching hit rate was about 3% higher in the rush hour, confirming that the vehicle''s movement was related to social times such as leaving and leaving work. The number of previous coordinate values of the RSU used by the vehicle that should be considered to predict the direction of movement could be derived through experiments. Depending on the network state, the amount of caching data received by about 10% may be increased by a flexible pre-caching operation method. The pre-caching hit rate through movement prediction derived a caching hit rate of about 73.4% for a single RSU and a caching hit rate of about 78.6% according to the operation method of the proposed technique.