| Underwater Acoustic Sensor Network(UASN)is a special network form deployed in Underwater Wireless Sensor Network(WSN).Underwater acoustic sensor nodes work together through acoustic communication,which can be used to monitor and collect data in the underwater environment.However,since underwater acoustic sensor nodes usually use battery power and work autonomously in complex Marine environment for a long time,it is difficult to renew the energy of underwater acoustic sensor nodes.In addition,unlike WSN,UASN adopts acoustic communication,so the existing routing protocols based on WSN are not well applicable to UASN.In UASN,routing protocol directly affects the energy consumption of nodes.Therefore,with the continuous development of modern network communication technology,designing a routing optimization scheme to improve the energy processing efficiency of UASN network has gradually become a hot topic in UASN field.This thesis studies the clustering routing problem in UASN,mainly including:This thesis introduces the UASN field,expounds the UASN network architecture,underwater acoustic channel characteristics and routing protocol classification,and explains the current research status of UASN network energy consumption optimization at home and abroad.In addition,This thesis aimed at the present situation of UASN topology clustering method on how to improve the energy efficiency between sensor nodes,the UASN based on energy equilibrium layer clustering algorithm,this algorithm through the depth of the underwater sensor nodes and energy of sensor nodes into clusters of different size,on the choice of cluster heads at the same time,comprehensive factors such as the residual energy of nodes,distance,weight the biggest node to become cluster head nodes are selected,made of the selected cluster head nodes more robust,and the selection of cluster heads is more reasonable.Simulation experiments show that the proposed algorithm can effectively alleviate the "hot spot" problem in the network,and can more effectively balance the energy consumption compared with the other two typical clustering algorithms in UASN.Aiming at the shortcomings of routing optimization algorithms in UASN networks,this thesis proposes an inter-cluster routing optimization algorithm based on reinforcement learning.In this algorithm,an intelligent routing reinforcement learning method is used to learn and filter data transmission paths between nodes.First of all,the proposed algorithm uses a reward function to assign a higher reward value to the path that consumes less energy.Secondly,the algorithm in the choice of path between the nodes,through the use of a kind of ε-greedy strategy to balance route choice,this strategy can make the network node in the selection of high reward values on the basis of the transmission path,still maintain a certain probability of randomly chosen path,the purpose is to better balance the ratio between the greedy and randomness,make the whole network automatically jump out of local optimum,which can accelerate the convergence speed of the network at the same time.Finally,the experimental results show that the proposed algorithm is effective in reducing UASN energy consumption and prolonging network service life. |
