Research On Energy-saving Routing Algorithm In WSN

Wireless Sensor Networks (WSN) face the problem that how to useenergy-limited battery to prolong the service life of the sensor application system asfar as possible. First of all, this paper gives an overview of WSN, and then it analyzesthe research situation and development trend of routing algorithm in WSN fromenergy-saving point and gives an intuitive understanding of the existing routingalgorithm. Finally, two algorithms are proposed, one is improved algorithm of LowEnergy Adaptive Clustering Hierarchy (LEACH) protocol based on Sliding Windowand Dynamic Number of nodes, the other is Uniform Structure and Load Balance Treerouting protocol, they are abbreviated as LEACH-SWDN and USLBT, respectively.Firstly, this paper analyzes the classical clustering algorithm LEACH and its twoimproved algorithms Low Energy Adaptive Clustering Hierarchy with DeterministicCluster-Head Selection (LEACH-DCHS) and Advanced LEACH (ALEACH), itanalyzes a result that LEACH hasn’t taken the aspect of residual energy informationinto account in set-up phase of clusters. At the same time, the cluster headsexpectation of energy-saving clustering algorithm LEACH exist optimal range, butthe improved algorithms did not meet the condition. So, the residual energy of thenodes is introduced to optimize the cluster-head selection, and a sliding window is setup to adjust the electing probability and keep stable the expected number of thecluster heads using the average energy information. Meanwhile, adjust the currentnumber of survival nodes to optimize the network performance dynamically, and thengive the complete LEACH-SWDN protocol. In order to set up the sliding window, aapproximate average energy information of all nodes was first introduced todynamically change the upper limit of the sliding window, and it gives a proof that thenumber of cluster head close to the optimal range in condition of the residual energyis introduced into the threshold formula that controls the cluster-head selection. Then,this paper optimizes the defines of sliding window by introduce the average energy ofthe nodes that have not already been cluster heads within current cycle to control theupper limit of the sliding window, and gives another proof that the optimizedLEACH-SWDN protocol meet the optimal expectation range completely. To compare the performance of LEACH-SWDN algorithm with the prevalentones this paper carries out simulations to measure the following metrics: networklifetime, number of cluster heads, energy efficiency and throughput. The simulationresults show that LEACH-SWDN optimizes the cluster-head selection as expected,and the First Node Dies (FND) value is improved in condition of the node with moreresidual energy has a bigger probability to be a cluster head. That is to extend thenetwork lifetime. It stabilizes the number of cluster head through the regulation of thesliding window, and solves the problem that the number of cluster heads is not stableby introducing energy information in the improved protocol such as LEACH-DCHSand ALEACH. This algorithm also improves the energy efficiency and the networkthroughput when the network lifetime has been prolonged.This paper analyzes the adjustment rules about balanced tree in AdjustableConvergecast Tree (ACT) protocol, and summarizes the feature that ACT can balancethe energy consumption of network nodes only in the transverse direction, it is to saythat the nodes that have the equal hops to the base station have the equal energyconsumption, and the unequal hops have large energy consumption gap.Simultaneously, the adjustment rule can not get a balanced tree after using the treeadjustment. Taking into account the deficiencies of the existing tree shape and clustershape routing protocol, this paper gives clustering method based grid and transmissionmodel based probability of depth, and accordingly gives USLBT protocol. Thisprotocol is a combination of cluster and tree structure, it constructs a routing structurethat has characteristics includes global balance and uniform structure between theclusters. USLBT algorithm introduces the clustering method based grid to divide thesensor deployment area into uniform cluster structure, and each cluster is regarded asa backbone node, all backbone nodes coalesces the balanced tree idea in ACT and itbalances the energy consumption of nodes that has equal hops to the base station inthe transverse direction. This algorithm introduces transmission model basedprobability of depth to avoid the emergence of death diffusion, and it balances theenergy consumption of nodes that in the same transmission path in longitudinaldirection. Preferable, the structural adjustment is confined within the backbone nodesin the maintenance phase of the routing topology, it avoids the problem that needs acentralized adjustment by the base station in ACT protocol, and the routing topologythat is adjusted in the maintenance phase still satisfy the global equilibrium in USLBTprotocol. To compare the performance of USLBT algorithm with the prevalent ones thispaper carries out simulations to measure the following metrics: network lifetime,completeness of data and structure of cluster. The simulation results show thatUSLBT balances the energy consumption among the network nodes and extends thefirst node dead time. It uses the stable and balanced topology to ensure thecompleteness of the monitoring data, so it also improves the energy efficiency.