Efficient resource allocation based on semi persistent scheduling sharing sensed information in C-V2X service

Abstract

The importance of standards in vehicular communication is increasing as the development and distribution of autonomous vehicles are in process. Accordingly, 3GPP standardized sidelink mode 4 in which the vehicle allocates resources by itself. In cellular V2X mode 4, a candidate single-subframe resource (CSR) excluding resources that other vehicles have reserved or already selected is created through sensing and sharing information, and when the number of resources in the CSR is more than 20 percent of the total resources, a resource with the minimum signal strength is randomly selected to transmit the packet. In this process, the information shared between vehicles is limited, so the amount of resources selected based on the sensing result may not be sufficient. Also if two or more vehicles select the same resource, a message collision occurs, which may threaten the driver's life or interfere with the traffic flow. The amount of vehicle's information is also limited, so the amount of resources selected based on sensing results may not be sufficient. In this thesis, to solve such a problem, we propose a method that can reduce message collisions and increase the amount of available resource by sharing or comparing sensing information measured by each vehicle and reselection parameters. The resource with largest difference between the shared sensing information and the sensing result of the resource candidate group has a high priority. So message collisions are decreased by reducing the probability of selecting the overlapped resource between vehicles. In addition, by sharing a reselection counter, which is the number of remaining transmissions of the vehicles, the available resources after transmission is finished during sensing are also included in the candidate group to increase the amount of available resources. In addition, in this thesis, the performance analysis of the proposed method is conducted using the urban driving simulation framework with the mobility of the vehicle. A number of vehicles send and receive messages in process of simulation, and a packet reception success rate is measured, which is proportional to the amount of sensing information received. Compared with the conventional method, it was confirmed that the amount of sensing information received when the distance between vehicles increased, the packet reception success rate was improved compared to the conventional method, and the amount of available resources increased by maintaining a high channel busy rate.