| This study focuses on the design of a geographical location routing protocol based on en-counter history and prediction in VMONET(Vehicular Mobile Opportunistic Network).The fixed geographic location and shape of traffic roads in urban environments allows for pre-dictable vehicle node trajectories,predicting in advance whether a vehicle node is at an intersection,which in turn determines the direction of packet delivery and the location of the next hop relay node.By making predictions to optimize the packet delivery path and bypass unnecessary vehicle nodes in the delivery process,the average delay is reduced and the stability of the connection between communication nodes is enhanced to adapt to the road characteristics of the urban environment and the rapidly changing VMONET.The GRPE(Geographical Routing Protocol Based on Prediction and Encounter History)routing algorithm designed in this study uses message survival time and buffer size as inde-pendent variables,respectively,to further analyze the performance of the routing algorithm on the three evaluation metrics of package delivery probability,average latency and routing overhead.The comparison experiments compare the GRPE routing algorithm designed in this study with the breakthrough algorithms GROOV(Geographic ROuting Over VANETs),PGR(Prediction-based Geographic Routing)and classical VMONET algorithms such as Epidemic,SprayAndWait and Prophet.In terms of package delivery probability and average latency,the simulation and Interpre-tation of the results show that the GRPE algorithm performs well than other routing algo-rithms,but has a higher routing overhead,which is related to the numerous impact factors that need to be calculated during its routing judgement.Overall,the GRPE routing algo-rithm designed in this study performs poorly in terms of routing overhead,but still within a manageable range,sacrificing routing overhead for a high message delivery probability and low average transmission latency. |