版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领
文档简介
1、Trading off Charging and Sensing for Stochastic Events Monitoring in WRSNsYu Sun*, Chi Lin*, Haipeng Dai, Qiang Lin, Lei Wang*, and Guowei Wu*School of Software Technology, Dalian University of Technology, Dalian 116023, ChinaState Key Laboratory for Novel Software Technology, Nanjing University, Na
2、njing 210024, ChinaDalian University of Science and Technology, Dalian 116052, China目录CONTENTSABSTRACTBackground:Wireless Rechargeable Sensor Networks (WRSNs)Our Work:Charging scheduling scheme optimization :Optimizing charging utilityTrading off charging and sensingReducing the performance lossAvoi
3、d the serious consequencesConclusions:Simulation: the performance of the proposed scheme is 19.7% higher than baseline algorithms;Realistic experiments indicates the feasibility and superiority in real scenes.Ideal AssumptionsProblem:HardwareConstraints ConflictsCharging ExclusivityPerformance LossM
4、otivation目录CONTENTSCONTENTSPart 01BackgroundPart 03SolutionPart 02ModelPart 04Theoretical AnalysisPart 05ExperimentsBackgroundPART 01 Background Model Solution Analysis ExperimentsLimited batteriesMaintenance difficultiesDeployment difficultiesLimited applicationEnergy bottleneck of traditional Inte
5、rnet of ThingsWireless Power TransferWireless charging technology realizes the remote and reliable supply of large amounts of electric energy, which provides a new solution to solve the energy bottleneck problem of wireless sensor networks.On this basis, Wireless Rechargeable Sensor Network is gener
6、ated. Background Model Solution Analysis ExperimentsWireless Rechargeable Sensor NetworksWireless Sensor NetworkWireless Power TransferWireless Rechargeable Sensor NetworkRechargeable SensorsSuper capacitorSensing surrounding environmentWireless Charging Vehicle (WCV)Equipped with wireless chargerch
7、arging panic sensorsAt present, the research of WRSN is mainly based on simulation environment. Background Model Solution Analysis ExperimentsIn the actual scene experiment, the phenomenon often occurs that the sensor node cannot work when being charged.Experimental phenomenaExisting problem: Chargi
8、ng ExclusivityThe charge and discharge states of supercapacitors cannot co-exist.The sensor node has simple structure and no power control circuit.Charging and sensing cannot be conducted at the same time.Sensor being charged Sensing is suspended Important events may be missed Serious consequences a
9、re caused.Main circuit structure of sensor node energy capture module Background Model Solution Analysis ExperimentsChallengesMotivationSimulations ignored charging exclusivity.Existing charging scheduling is unreasonable.Possibly causing serious consequences.A serious conflict between experimental
10、phenomena and simulation assumptions.We propose a scheme which is not only suitable for simulation theoretical analysis, but also has excellent performance in practical applications. Background Model Solution Analysis ExperimentsIt is non-trivial to quantitatively leverage the utility gain vs. utili
11、ty loss yielded by the charging behavior of WCV.The solution space of solving this problem is infinite.Constructing traveling path of WCV is equivalent to solving a TSP.The objective function of this problem is non-linear and hard to analyze its properties. Background Model Solution Analysis Experim
12、entsContributionsModelPART 02 Background Model Solution Analysis ExperimentsNetwork ElementsSymbolsBehaviorSensorsMonitor stochastic events at all PoIs within its sensing rangePoints of Interests (PoIs)Stochastic events are generated at PoIsSojourn spots for WCVTravels within the network, and stops
13、at sojourn spots to charge surrounding sensors Background Model Solution Analysis ExperimentsNetwork modelEvent model(due to charging exclusivity) Background Model Solution Analysis Experiments(due to charging exclusivity)Before charging:After charging:Monitoring utility computation Background Model
14、 Solution Analysis Experiments(due to charging exclusivity)Energy consumption of WCVCharging cost:(Consumption of charging sensors)Traveling cost:(Consumption of traveling)The power transmission model is as follows:whereCharging model Background Model Solution Analysis ExperimentsProblem Formulation
15、Charging Exclusivity Optimization(CEO)ProblemHow to select the appropriate sojourn spots set X for WCV to charge surrounding sensors such that the total charging utility of the network under charging exclusivity is maximized.The CEO problem is the coupling of multiple challenging problems that has h
16、igh computational complexity.Difficulty AnalysisInfinite solution spaceNon-intuitive natureIntroduction of TSP Background Model Solution Analysis ExperimentsSolutionPART 03 Background Model Solution Analysis ExperimentsArea Distretization: Circular DiscretizationArea discretization: (a) Charging pow
17、er discretization; (b) Draw concentric circlesIn order to reduce infinite solution space into a finite set, the continuous network area is divided into several subareas through area discretization:Discretization Process:Piecewise constant functionDraw concentric circlesUniform within each subarea Ba
18、ckground Model Solution Analysis ExperimentsArea Distretization: Polygonal DiscretizationReplacing each concentric circle with an inscribed regular polygon inside it.The curved polygon subareas formed by circular discretization brings difficulties to the following path planning problems. We propose
19、polygonal discretization to modify them: Background Model Solution Analysis ExperimentsConvex Polygonal DiscretizationWith the aforementioned parameter settings, the approximation error of area discretization is bounded to:The polygonal subareas are not all convex areas, further modification:Dividin
20、g subareas into convex polygons Background Model Solution Analysis ExperimentsWhen the solution space is reduced from infinite set to finite set, CEO problem can be transformed into CEO-R problem:The transformed CEO-R problem is a nonlinear combinatorial optimization problem, which is still difficul
21、t to solve. Background Model Solution Analysis ExperimentsProblem ReformulationTouring among convex polygonsCEO ProblemTSP among spotsTraveling pathCEO-R ProblemTouring among areas1 X. Tan and B. Jiang, “Efficient algorithms for touring a sequence of convex polygons and related problems,” in TAMC, 2
22、017, pp. 614627.Circular DiscretizationPolygonal DiscretizationConvex PolygonsTouring among convex polygon areas1(non-convexNP-Hard ) Background Model Solution Analysis ExperimentsTraveling Path Constructionof WCVTraveling pathof WCVTo solve the CEO-R problem, we propose an efficient approximation a
23、lgorithm:Considering area selection and path construction at the same time.Different selection lead to different traveling path.01Cost-Benefit Ratio is utilized to evaluate sojourn areas.Cost-Benefit Ratio is defined as the ratio of marginal benefit growth and marginal cost growth. Iterative selecti
24、on according to greedy strategy.02When Cost-Benefit Ratio is negative, selection terminated.Negative Cost-Benefit Ratio indicates no marginal benefit growth.03 Background Model Solution Analysis ExperimentsApproximation AlgorithmSelect until budget is exceeded.For each newly selected area, recalcula
25、te corresponding charging utility and WCV traveling path.04Select suboptimal solution when the optimal one exceeds the budget.When any selection will exceed the budget, jump out of the iteration.05Compare with the baseline value.Output set is compared to the set with a single element which has the m
26、aximum marginal benefit growth. The one with larger utility is output as the result set.06 Background Model Solution Analysis ExperimentsApproximation AlgorithmTheoretical AnalysisPART 04 Background Model Solution Analysis ExperimentsNP HardnessCEO ProblemSelect elements with maximum benefit under b
27、udget constraintsIgnore the path planningBudgeted maximum coverage problemcan be reduced toNP-HardSimplifiedNP-HardNP-HardTo solve the CEO problem, general algorithms cannot obtain the optimal solution in acceptable time, so it is necessary to reduce the difficulty of solving the problem through pro
28、blem transformation and approximation algorithm. Background Model Solution Analysis ExperimentsNonnegativity Background Model Solution Analysis ExperimentsProperties of objective functionMonotonicityProbability gain and probability loss in different time periods.Transitivity of monotonicity and subm
29、odularity Background Model Solution Analysis ExperimentsSubmodularityPossible relationships of charged, to be charged, and uncharged sensor setsTo prove the submodularity, the four possible relationships among charged, to-be-charged and uncharged sensor sets in two different situations A and B are c
30、lassified, and the marginal benefit growth of newly added elements in these cases are analyzed.Specific charging processBy analyzing the specific charging process under various relationships, the charging utility increment under the conditions A and B is quantitatively compared, and the submodularit
31、y is proved. Background Model Solution Analysis ExperimentsExperimentsPART 05 Background Model Solution Analysis ExperimentsParameterValueNetwork size100m*100m4010050J10000J10154J/m0.5J/s1010m1.5m/sSimulation parametersIn this experiment, we simulate the effect of charging exclusivity, and set senso
32、rs to stop working until the charging is completed.Simulation SetupBaseline SetupME: Maximizing energy received by sensors;CHASE1: State-of-the-art algorithm;CEO: Our scheme;CCO: Our scheme in ideal simulation assumption (without charging exclusivity);CHASE-C: CHASE without charging exclusivity. Bac
33、kground Model Solution Analysis Experiments1 H. Dai, Q. Ma, X. Wu, G. Chen, D. K. Y. Yau, S. Tang, X. Li, and C. Tian, “CHASE: charging and scheduling scheme for stochastic event capture in wireless rechargeable sensor networks,” IEEE Transactions on Mobile Computing, vol. 19, no. 1, pp. 4459, 2020.
34、Under different parameter settings, the performance of our scheme is 21.3% higher than other baseline algorithms on average.Charging utility in different settings Background Model Solution Analysis ExperimentsIn the actual scene experiment, the WRSN is deployed by simulating the fire monitoring network.10 monitoring points are set in different positions in an open area, which are regarded as PoIs, 25 sensors are deployed around these PoI
温馨提示
- 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
- 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
- 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
- 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
- 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
- 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
- 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
最新文档
- 沈从文《街》课件
- 银龙溪两岸护坡及堤顶道路工程施工组织设计方案
- 年产2000台玉米收获机技术改造项目可行性研究报告
- 法律保护我们的人格尊严课件
- 2015年重庆市B卷中考满分作文《我们携手走进友谊》
- 《条件随机场CRF》课件
- 《成长的烦恼》作文讲评课件
- 展览中心铝塑板安装施工协议
- 科技论文写作讲座课件
- 城市安全建设项目立项指南
- 2024(新高考2卷)英语试题详解解析 课件
- 《天气学原理》考试复习题库(含答案)
- 大庆2024年黑龙江大庆市龙凤区人才引进80人笔试历年典型考题及考点附答案解析
- 烟酒行转让合同范本
- 2024年高考数学模拟试卷附答案解析
- 荆楚民艺智慧树知到期末考试答案章节答案2024年湖北第二师范学院
- 穿脱隔离衣的流程及注意事项
- 外国文学智慧树知到期末考试答案章节答案2024年九江职业大学
- 拼多多营销总结报告
- 电子信息类专业《计算机网络》课程教学的改革与实践
- 钢板加固梁施工方案
评论
0/150
提交评论