Distributed Optimal Rate-Reliability-Lifetime Tradeoff In Wireless Sensor Networks

Wireless sensor networks (WSNs) are Resource-constrained wireless networks,which usually deployed in harsh environmental conditions. Sensor node has poorcommunication performance. Its transmission power is limited. It is difficult for theperception information pass back to the sink node accurately and reliably. Therefore,how to use the limited resources of the sensor nodes reliable and high-speedtransmission sensing information is a hot and difficult research in WSNs. In theapplications of WSNs, a high data rate can be obtained on a link at the expense oflower delivery reliability, which results in more energy consumption and a reductionof network lifetime. Therefore, there is an inherent tradeoff among the informationdata rate, reliability and network lifetime. Hence, how to provide higher networkthroughput, meanwhile, ensure the reliability of information and prolong the networklifetime become a significant research topic in WSNs.Based on the full analysis the research status at home and abroad of rate,reliability and network lifetime for WSNs, we do the following research andinnovation works:(1) We first proposed the rate-reliability-lifetime tradeoff problem in energy-constrained WSNs with link capacity constraint, reliability constraint and energyconstraint. We introduce the weight parameters, which combine the objectives at rate,reliability, and lifetime into a single objective to characterize the tradeoff among them.We can achieve the tradeoff optimization among the three objectives by adjusting theweight parameter.(2) We address the rate-reliability-lifetime tradeoff problem in integratedreliability policy. The objective function is decoupled by introducing auxiliaryvariables. We introduce the concept of inconsistent coordination price to balance theenergy consumption of the sensor nodes. We adopt subgradient dual decompositionmethod to solve the optimization problem, and develop the distributed algorithm based on integrated reliability policy (DAIRP). We validate the algorithm throughMatlab simulation and analyze the performances of the algorithm. The simulationresults show that: the distributed algorithm proposed in this paper can increasenetwork throughput, meanwhile, ensure the reliability of data transmission andprolong the network lifetime.(3) We address the rate-reliability-lifetime tradeoff problem in differentiatedreliability policy. As each link can provide a different code rate to each of the nodesusing this link, it leads to the optimization formulation of the tradeoff problem isneither separable nor convex. Through a series of transformations, a separable andconvex problem is derived, and an efficient distributed algorithm based ondifferentiated reliability policy (DADRP) is proposed. Numerical examples confirmits convergence，and compare the performance of DAIRP and DADRP. Simulationresults show that: the convergence speed of the DADRP is slower than the DAIRP, butit can obtain more network utility. They are applied to different network scenarios.(4) We address the rate-reliability-lifetime tradeoff problem in randomtime-varying channel conditions. We established rate-reliability-lifetime collaborativeoptimization model under random time-varying channel conditions, and turn thisproblem into a separable convex problem. In practice, we do not have a prioriknowledge of the underlying probability distribution for the channel states. So wedevelop stochastic subgradient method to solve the problem. This algorithm does notrequire a priori knowledge of the underlying probability distribution for channel states.It only requires the information of the channel state at current iteration. We proves theconvergence of stochastic subgradient algorithm. We validate the algorithm throughMatlab simulation. Simulation results show that: stochastic subgradient algorithm isconvergent. It can obtain the global optimal solution for the rate-reliability-lifetimetradeoff problem.