Research on Handover of Urban Rail Transit Based on PER-DDPG Algorithm
In order to solve the problems of the traditional IEEE802.11 off-zone switching mode,such as high switching delay and ping-pong switching,a deep Q-Network off-zone switching algorithm is proposed.A crossover algorithm of deep Q-Network is proposed.By extracting and inputting the characteristic state information of train op-eration,a crossover switching model is constructed considering the dynamic information of train running speed,field strength and switching threshold.Meanwhile,Prioritized Experience Replay-Deep Deterministic Policy Gradi-ent is prioritized to the complexity and stability of the algorithm's time cost.The spatial information of train status is transmitted to PER-DDPG network for optimization analysis.The results show that the time calculation cost of the optimized train crossover model based on PER-DDPG algorithm is reduced.the packet transmission delay is re-duced by about 55%.
NETL
NSTL
万方数据
