Research On Resource Allocation Algorithm Based On Train-To-Train Communication In Urban Rail Transit

The Communication Based Train Control(CBTC)system is widely used in the field of urban rail transit.However,the open-band Wireless Local Area Network(WLAN)technology adopted by the CBTC system can support co-frequency interference.Problems such as poor mobility can no longer meet the requirements for safety and traffic efficiency in the rail transit field in the new era.Not only that,the existing CBTC system only supports Train-to-Ground(T2G)communication,which will cause the train to fail to obtain the status of the preceding train in time due to the failure of trackside equipment and cause a collision.Therefore,in order to improve the reliability and stability of train communication and reduce the complexity of trackside equipment,consider adding train-to-train(T2T)communication technology to the existing CBTC system.This paper mainly studies the application of D2 D communication technology in CBTC system.First,model the CBTC system based on D2 D communication technology,and then propose a resource allocation algorithm suitable for the system.The research results of this paper have a certain practicability and have certain reference significance for the development of T2T communication in the future.The specific research content of the paper is as follows:(1)Through the system modeling,this paper studies the joint communication mode selection and resource allocation when the T2 G communication mode and T2T communication mode coexist.On the premise of ensuring the signal-to-noise ratio of the communication link,the channel capacity of the entire system is maximized.First,a communication mode selection algorithm is proposed,the purpose of which is to enable the train to adaptively select a communication mode with better communication quality.Then,based on the choice of communication mode,a resource allocation algorithm based on the Lagrangian dual function method is proposed,through which the optimal transmission power of the train in different communication modes can be obtained.The simulation results show that compared with other resource allocation algorithms,the algorithm proposed in this paper can significantly improve the channel capacity of the system.(2)According to the system model proposed in this paper,the T2T communication spectrum sharing problem is studied.The T2 G communication link can be reused by the T2T communication link.In order to reduce the interference caused by reuse,the spectrum sharing problem is modeled as a Multi-Agent Reinforcement Learning(MARL)problem.Deep Q learning relies on the experience pool,but due to the instability of the experience pool,the commonly used MARL algorithm is not compatible with the experience pool.Therefore,this paper adopts the Fingerprint-Based Deep Q-network(FDQN).In this system,each T2T communication link acts as an agent,and multiple agents cooperate with each other in a distributed manner.The environment interactively updates the network parameters,and selects the best spectrum and transmission power based on the learning experience.The simulation results show that through distributed coordination,the channel capacity of the system can be greatly improved and the probability of successful T2T communication data transmission can be improved.