Research On Load Balanced Routing Protocol For Underwater Acoustic Wireless Sensor Networks

With the rapid development of the marine economy,the ocean plays an important role in the production and life of human beings.For the purpose of environmental protection and resource extraction,the demand for monitoring and detecting of the marine environment and marine resources is increasingly urgent.Underwater acoustic wireless sensor networks(UAWSN)provide a new solution for automated monitoring of marine environment through environmental awareness,autonomous networking,and collaborative transmission.UAWSN have become an important part of marine science and technology.In many research problems of underwater acoustic wireless sensor networks,routing protocol is used in important processes such as the formation of networks and collaborative transmission of sensor data.It is one of the key technologies to realize the connection and data transmission of underwater acoustic wireless sensor networks.Due to the special physical environment of the ocean,the routing protocol of the underwater acoustic wireless sensor networks needs to deal with the challenges of small communication bandwidth,limited energy of underwater acoustic nodes,and long communication delay.In order to deal with the above challenges,it's critical to effectively utilize the limited communication bandwidth,balance the overall energy consumption of the networks,and prolong the networks survival time by reasonably distributing the load in the routing protocol,and it's related to multiple network performance indicators.By focusing on the characteristics and challenges of underwater acoustic wireless sensor networks and the requirements of load distribution in the networks,this thesis researched the routing protocol of underwater acoustic wireless sensor networks based on load balancing in various scenarios.Firstly,for the problems of limited bandwidth,limited energy and difficulty of obtaining global information in UAWSN,this thesis proposed a distributed underwater load balancing routing protocol based on reinforcement learning in a static network.Modeling the load distribution problem of nodes as a reinforcement learning problem.Through the learning of historical network state information,nodes have distributed load distribution decision-making ability,which can effectively reduce information interaction.At the same time,in order to save energy and accelerate convergence,the algorithm introduced evolutionary game theory to optimize the exploration strategy in the learning algorithm.Then,this thesis proposed an underwater dynamic network routing protocol based on reinforcement learning in a dynamic network for the moving nodes in the underwater.Modeling state space for reinforcement learning through shared attributes,and constructing a variable-sized state group to cope with the dynamically changing surrounding topology.For the variable-sized state group,designing a probability-based exploration strategy for reinforcement learning.Simulation experiment results show that the above algorithms proposed in this thesis effectively distributed network load in various scenarios,improved network transmission efficiency,optimized the balance of network energy consumption,and prolonged network survival time.Finally,this thesis designed the underwater acoustic communication software system based on Linux for the needs of the actual deployment of the underwater acoustic wireless sensor network,and verified the integrity of its communication function and the correctness of the protocol logic through experiments.The research results of this thesis provide a reference for the subsequent research and deployment of underwater acoustic wireless sensor networks.