Research On Location Based Routing Protocol For Vehicular Ad-hoc Network

Vehicular ad-hoc network(VANET)not only has a wide range of application prospects in ensuring traffic safety and solving traffic congestion,but also play an important role in intelligent transportation system(ITS).Vehicle obtains information about the surrounding road conditions through communicating with other vehicles and roadside units..In this way VANET builds a real-time traffic information exchange system from vehicles to vehicles(V2V),vehicles to people(V2P),and vehicles to infrastructure(V2I).High mobility is the most typical challenge of VANET.It will lead to frequent topology changes,disconnected communication links,increased end to end delay and eventually damage the robustness of the whole system.As a key technology to solve these problems,routing protocols have received extensive attention from scholars at home and abroad.In this paper,we propose an improved location-based routing protocol and a reinforcement learning-based routing protocol for VANET.These protocols can reduce the end to end delay and improve the packet delivery ratio in high-speed VANET environments.Traditional routing protocols only consider the distance factor,but in this thesis we propose an enhanced-location-based routing protocol(ELRP)by taking into account the effects of distance,relative velocity,node density,node travel direction,etc.The proposed protocol can better adapt to both high-speed and low-speed VANET environments.At the same time,the forwarding rules are redesigned to reduce unnecessary forwarding.By introducing adaptive broadcast interval,the broadcast storm problem is solved,the routing overhead and overall network burden are reduced.Location prediction before forwarding can be better applied to high-speed vehicular network environment and increase transmission success rate.The proposed ELRP only selects next hop based on the current local state of the system and does not fully consider the influence of the global state of the whole system.To further improve the performance of the protocol,this paper proposes a dynamic adaptive reinforcement learning-based routing protocol.By considering several link evaluation metrics,a new forwarding reward strategy is proposed.On the one hand,the proposed protocol takes the effects of the distance,the link duration time,and the direction of movement into consideration.On the other hand,the protocol introduces a heuristic learning approach to increase the exploration space and accelerate the convergence of learning.Reinforcement Learning approach enables vehicle nodes to learn from the system’s historical state,which can further improve the network transmission performance.This protocol is able to adapt to both dense and sparse vehicle environment,high mobility or low mobility scenarios and therefore can perform well in VANET.The performance of the proposed protocols is verified by simulation using NS-3 and SUMO.The simulation results show that both proposed protocols achieve higher packet delivery ratio,lower end to end delay,less routing overhead and higher throughput than the traditional protocols.This indicates that the proposed protocols in this paper can be better adapted to the VANET environment and have higher reliability and stability.