Abstract: In the road traffic network,traffic congestion problem is a complicated dynamic process of interaction between flow and the structure of the network.Through the vehicle path planning,the integration of the road network grid scheduling is realized,and traffic throughput can be improved.The traditional method adopts parallel microscopic traffic dynamic prediction algorithm to realize the vehicle congestion scheduling and vehicle routing planning,but the algorithm can not accurately judge the density of vehicles,and the performance is not good.An improved anti congestion vehicle path planning algorithm was proposed based on cloud grid integrated scheduling.The cloud road network model is constructed based on Small-World model,and RFID label is used to collect the traffic information.The intrinsic mode function weighted average is used to calculate the vehicle congestion state function of each lane,and the density of vehicles in all lanes is obtained from the statistical average available vehicle density cluster.The traffic road network congestion detection algorithm was designed,searching for the current road information of individual one-dimensional neighbor,then the vehicle path planning and best objective function optimization are realized.The dynamic game way is used to get the approximate optimal trajectory to improve the path planning algorithm.The simulation results show that the algorithm can accurately achieve the optimal vehicle path planning and control,and traffic speed and network throughput performance are improved in severe congestion state.It has better performance than traditional method.

Key words: Cloud grid,Road network model,Throughput,Path planning