Research On Routing Algorithm Of Underwater Sensor Network For Void Hole

With the continuous exploration and development of marine resources,the application scene of underwater wireless sensor network(UWSNs)has been gradually expanded.In sparse UWSNs with low node density,there are many problems,such as frequent routing holes,high network energy consumption,low success rate of packet delivery and so on.How to overcome the problems existing in underwater environment,such as low bandwidth,high delay,high dynamic environment,limited node energy and so on,how to improve the performance of UWSNs network through routing algorithm is an urgent problem to be solved.Traditional UWSNs routing algorithms often lack a grasp of the overall energy distribution of the network,and have the problems of uneven energy distribution and short network life.with the continuous development and progress of sensor technology and the continuous improvement of the related theories of intelligent algorithms,it is very promising to introduce reinforcement learning technology into underwater routing problems to design intelligent algorithms suitable for application scenarios.Aiming at the problems of high dynamic and uneven energy distribution in sparse dynamic underwater wireless sensor networks,this paper proposes an energy efficient routing algorithm(EERFQ)based on reinforcement learning and message feedback mechanism.The research content of this paper mainly includes the following aspects:(1)aiming at the common routing holes in sparse dynamic networks,an empty node avoidance strategy based on node candidate factor is proposed.(2)aiming at the problems of packet loss and low success rate of packet delivery caused by routing holes,a packet recovery and retransmission mechanism based on message feedback is designed to recover the effective data of empty nodes and improve the success rate of delivery.(3)comprehensively considering the energy factors and depth information of network nodes,and combining with node candidate factors,an energy-efficient routing algorithm based on reinforcement learning is proposed,and the algorithm flow and reward function are designed.it makes the algorithm better adapt to sparse networks on the premise of maintaining a certain routing performance.(4)aiming at the redundant forwarding problem in the routing process,a suppression strategy is designed to save network communication resources and reduce network energy consumption.This paper simulates the EERFQ algorithm through the network simulation platform NS-3,including the network performance in different application scenarios finally,and compares the network performance with several other typical algorithms in the same scenario,and analyzes the simulation results in detail.Simulation results show that EERFQ algorithm has some problems such as high energy consumption and long average forwarding delay in underwater wireless sensor networks with high node density,but in sparse networks,its performance is significantly better than other algorithms,and it can adapt to dynamic networks,effectively deal with routing holes,balance network residual energy distribution and prolong network lifetime.