| The cellular vehicle-to-everything（C-V2X）technology can support ultra-reliable and low latency communications（URLLC）scenarios in 5G,which is one of the key technologies for highly automated driving in the future.The sensing based semi-persistent scheduling（SPS）algorithm is used in C-V2 X Mode 4 to process the basic messages of security applications,allowing the vehicle to independently select radio resources and communicate directly through the PC5 interface.Therefore,Mode 4 is considered to be the most important communication mode for C-V2 X communication.Since Mode 4 is prone to packet transmission collision and unreliable communication in a dynamic environment,and the network congestion caused by high traffic load will affect the throughput,the radio resource scheduling in mode 4 will remain a challenge.In this paper,radio resource scheduling is studied based on sensing based semi-persistent scheduling and reinforcement learning theory,and distributed congestion control is studied based on transmission power control and non-cooperative game theory.The specific innovations are as follows:（1）For the problem of packet collision in Mode 4 with high mobility,a reservation and reuse combined Q-learning semi-persistent scheduling（RRC-QSPS）algorithm is proposed.Firstly,this paper analyzes the packet collision process in dynamic C-V2 X network and establishes the theoretical model of packet collision.According to the theoretical model,the effects of reservation probability and reuse times on packet collision probability in high-density channel environment are analyzed;Then the instantaneous return function of RRC-QSPS algorithm is designed according to the packet collision.The proposed algorithm allows the vehicle intelligent decision-making to reserve and reuse radio resources in the dynamic network environment,so that the selected resource can adapt to the dynamic changes of the network.（2）For the problem of limited throughput in Mode 4 with high load,a distributed congestion control mechanism using non-cooperative game based transmit power control（NG-TPC）algorithm is proposed.Firstly,this paper studies the limitations of traditional distributed congestion control in Mode 4 and establishes the theoretical model of data transmission ratio based on drop packet transmission（DPT）,and analyzes the impact of DPT on system throughput;Then this paper constructs the throughput problem as a non-cooperative game model,designs the utility function and its price factor,proves the existence and uniqueness of Nash equilibrium,and uses the hybrid particle swarm optimization algorithm to solve the optimal solution of the game model.The proposed algorithm equalizes the channel interference and data transmission rate,and alleviates the limitation of throughput.Finally,three dynamic scenarios of traffic density,traffic load and traffic speed are simulated in LTEV2 Vsim simulator,then the comparison simulation and experiments of network performance are carried out.The simulation results show that in the dynamic traffic load scenario,compared with the resource location information based SPS algorithm,the RRC-QSPS algorithm improves the packet reception ratio by 7.1% and reduces the packet update delay by 10.1%.In the high load scenario,compared with the adaptive transmit power control algorithm,the NG-TPC algorithm improves the average throughput and packet reception ratio by 6.2% and 6.8% respectively.When DPT is enabled,the NG-TPC algorithm improves the average throughput and packet reception ratio by 8.5% and 9.3% respectively.Experiments verify the effectiveness of the innovative algorithm proposed in this paper.The new method can still maintain high reliable and low latency communication quality in high-speed mobile and high load scenarios,and has a good application prospect. |
