Design Of Routing Protocol Based On Clustering For Fishery Ad Hoc Networks

With the rapid development of China’s marine economy,in order to meet the increasing demand for marine communication,we need to continuously improve the overall performance of the marine communication system.However,the characteristics of Ad Hoc networks,such as centerless,self-organized and independent of fixed infrastructure,are consistent with the characteristics of marine communication environment.Therefore,the wireless Ad Hoc networks have been applied in the field of marine communication,and gradually become a research focus in this field.As the core technology in Ad Hoc networks,routing algorithm determines the overall performance of the network.Based on the characteristics of fishery Ad Hoc networks and the actual application scenarios,routing protocols suitable for marine fishery communication systems are proposed in this thesis,and the performance evaluation is accomplished by using NS2(Network Simulator Version 2).In order to reduce the collision of information between fishing vessels in the ports of fishery communication system,a routing protocol based on cellular and time slot named RPCT is proposed in this thesis.In RPCT protocol,time frames are allocated in combination with geographic location in the phase of neighbor nodes discovery.Nodes send HELLO messages through competing time slots in corresponding time frames,thus reducing message collisions in the phase of neighbor nodes discovery.Q learning is introduced in the process of cluster head election to improve the stability between cluster head and cluster member nodes.By simplifying TC(Topology Control)message and forwarding it through cluster heads,the routing overhead is reduced and the topology discovery efficiency is improved.In addition,a greedy forwarding strategy based on the geographic location of the cluster heads is adopted,which shortens the time of route establishment and reduces the maintenance overhead between routes.Simulation on NS2 platform shows that compared with OLSR(Optimized Link State Routing)protocol and AODV(Ad Hoc On-Demand Distance Vector Routing)protocol,the proposed RPCT protocol has smaller average end-toend delay and higher packet delivery rate and throughput.In addition,in order to improve the communication quality and provide stable data transmission between fishing vessels,a routing protocol based on cellular and reinforcement learning named PCRL is proposed.In PCRL protocol,cluster is taken as a unit for Q learning.After periodic broadcast,calculation and update of Q tables by cluster heads,the optimal paths between clusters can be obtained.This protocol uses the idea of on-demand route discovery.When there is a route request,the cluster heads are responsible for forwarding RREQ(Route Request)messages to reduce network overhead.In the process of routing decision,the destination node selects the optimal next hop cluster that reaches the source node’s cluster according to the Q table,and in the optimal cluster,the optimal next hop node is selected through the corresponding node selection strategy,that is to say,the route is established hop by hop through the dual decision-making on macro and micro level,so as to improve the stability of the data link.The simulation results of NS2 platform show that the PCRL protocol can provide higher packet delivery rate,network throughput and smaller endto-end delay than AODV and DSR(Dynamic Source Routing)protocols.Furthermore,the simulation results also show that compared with the RPCT protocol,the PCRL protocol can provide more stable and reliable data transmission services while carrying a large number of data transmission services under the condition that it is not sensitive to the delay requirements.