Research On Routing Methods In UASNs Based On Intelligent Algorithm

Underwater routing protocol is one of the most important aspects in Underwater Acoustic Sensor Networks(UASNs)and has an irreplaceable place in oceanic information technology.The complexity of ocean is the reason for the differences between routing protocols in Wireless Sensor Network(WSN)and UASNs.Routing protocols in UASNs based on intelligent algorithms are studied in this paper.The main contributions include:1.This paper studies the impact of ocean condition on the communication and networking of underwater sensors.Simulation setup for underwater sensor networks is built upon NS-3 simulation platform.And the design challenges and performance metrics of underwater routing algorithms are given.2.A fuzzy vector based forward routing protocol is proposed,and this broadcast method is suitable to 3-D dense underwater networks.The adaptive algorithm in vector based forward mechanism can effectively avoid the transmission of redundant data packets.Beyond that,the single-hop transmission is proved to be energy-efficient.In addition,a multi-parameters based routing decision method is proposed,of which 3 constraint factors: valid distance,projection and residual energy are inputs of fuzzy logic system.The proposed method can effectively extend the life span of networks and exhibit its decent performance in energy-efficiency and dynamic adaptability.3.The application of reinforcement learning in underwater routing protocol is studied,and this unicast method is suitable to 3-D sparse underwater networks and AUV networks.In 3-D sparse underwater networks,a collaborative reinforcement learning based on feedback underwater routing protocol is proposed.The nodes can adaptively modify the state transfer possibility function based on the feedback of link quality information.The utilization of node depth,residual energy and neighbor nodes information in reward function helps enhance the dynamic adaptability.In AUV networks,an AUV aided efficient routing protocol is studied.AUV shifts along the pre-defined routes,which causes low convergence speed of reinforcement learning.Thus,we introduce the backward propagation from AUV,and enable bidirectional information exchange which greatly boosts the convergence speed.This paper mainly researches on the intelligent routing algorithm for UASNs.Multiple routing protocols based on intelligent algorithms are proposed from different aspects,and effectiveness of algorithms is validated by simulations.