Energy Modernization Approach for Data Collection Maximization using Mobile Sink in WSNs

Transcription

1 Eergy Moderizatio Approach for Data Collectio Maximizatio usig Mobile Sik i WSNs Kumar Swamy B.V 1*, Dr. Gowramma Y.P 2, Aada Babu J 2 1 Robert Bosch Egieerig ad Busiess Solutios, Bagalore, Idia. 2 Kalpataru Istitute of Techology, Karataka, Idia. * Correspodig author Abstract I this paper a ovel approach of data collectio maximizatio usig mobile sik is cosidered for Moderizatio of eergy i wireless sesor etwork. Mobile sik improves the eergy of the system by travellig alog the path to collect data from oe-hop sesors periodically. Sesor are powered by eergy reewable resources ad depleted alog a path. Wheever we desig a etwork, time varyig characteristics of EH poses a great challege. So, i this paper we formulate a etwork that deals with multi-rate trasmissio ad parallel trasmissio time schedulig. A improved protocol is derived where ot oly cosiders the reewable of eergy, ode havig higher rate of reewable capacity will be erolled to participate more ito data trasmissio. Reliability of the etwork is studied ad is proved that QoS parameters shows better results with higher eergy expese. The proposed algorithms ad protocols are validated through simulatio experimets usig Network Simulator (NS2). Keywords: Eergy Cosumptio Rate (ECR), Eergy Reewable Rate (ERR), Eergy Harvestig (EH). vibratio eergy, wid eergy, etc. So, i the cocept of eergy harvestig, sesors recharge their batteries from their surroudigs. To sustai the etwork operatio i EH-WSN eergy cosumptio rate (ECR) should always be greater tha the eergy reewable rate otherwise sesor will lose their eergy. Data collectio i WSN usig MS havig fixed mobility has differet objectives as, INTRODUCTION Wireless Sesor Network has various applicatio i differet field such as i moitorig, eviromet parameter, health parameter ad i surveillace area icludig traffic parameter etc. Wireless sesor etwork cosists of thousads of tiy low cost ad low power devices kow as sesor odes (SNs), where each trasmittig ode i the etwork is resposible for [1] [2] collectig ad processig the data or iformatio ad trasmits to the MS or base statio. I traditioal sesor etwork [3] thousads of sesors are deployed i radom or predefied maer. Sesor odes sese the data ad trasmit to the static sik / BS through multi hop trasmissio., odes ear to sik suffer with heavy load ad deplete their eergy faster.a hole is created aroud the sik ad iformatio caot be routed to sik due to absece of data forwardig odes, which reduces the etwork performace. So, the sik mobility comes out with remedy to overcome eergy hole problem. For the improvemet [4][5] of etwork performace such as reducig the eergy dissipatio ad ehacig the etwork lifetime, sik mobility cocept has bee take. Most existig studies focused o the eergy cosumptio because sesors are powered by eergy limited batteries. So, there are various method to reduce the limitatio of battery such as, deployed sesor ca charge their battery from [7] [11] solar eergy, Figure 1. Data collectio i WSN usig mobile elemets with fixed mobility patters uder differet objectives I covetioal WSN for the data collectio maximizatio problem, a iteger liear programmig (ILP) [13] model is formulated. But i EH-WSN, each deployed sesor charges their batteries from the surroudigs, which poses time varyig characteristics. To respod time varyig ature traditioal ILP method takes too much time ad suffers from poor scalability ad will ot applicable due to quick chages i eergy profiles of sesors. So, i this paper, data collectio i EH-WSN is cosidered with mobile sik, which aims to formulate a optimizatio problem for fidig a optimal path for MS ad time schedulig for each sesors deployed at differet locatios to maximize the collected data, icludig multi-rate trasmissio ad slot schedulig of trasmissio time for each sesor. RELATED WORK This sectio briefly describes the related works carried out by researchers i the area of data collectio maximizatio i EH- WSN usig mobile sik. Due to advacemet i WSNs 6655

2 architecture based o mobile sik has bee studied by several researchers. Use of sik mobility [4][3][10] shows the improvemet i the etwork life time ad reduces the eergy cosumptio of sesors by balacig the workload amog the sesor. I geeral sik mobility is divided i three category : sik with fixed mobility which moves alog a pre-defied path, sik with radom mobility which is ofte mouted i aimals, humas ad sik with cotrolled mobility which is ofte used for data collectio. Frascesco [4] survey the data collectio i WSN thoroughly, where sik moves o the fixed trajectories, which foud may researchers attetio due to its practical applicatio i wide spread rage of its practical applicatio [3]. At differet time havig differet locatio, movemet of sik ca be determied by usig a LP method for the joit problem proposed by Wag. Usig a fixed mobility for data collectio from sigle hop sesor ode ot oly improves the etwork throughput but also reduces the data delivery latecy ad reduces the eergy cosumptio. For the etwork throughput poit of view i covetioal WSN the mai aim is to maximize the total collected data. Liag ad Juo [5] itroduced a joit optimizatio problem havig multiple mobile sik with h-hop costraied for statioary etwork. For this a three-stage heuristic is devised which cosists of calculatig time profile at each locatio ad improves the total collected data. With the recet advacemet i EH-wireless sesor etwork, sesors ca recharge their batteries from surroudigs. So, data collectio problem is ivestigated by X. Re, et al. [8] most of the works o data collectio usig a MS i EH-WSNs are defied with the objective of etwork throughput maximizatio [9], [10], [14]. Due to various applicatios for data collectio o the sigle or direct path, the authors i [9], [1] ivestigate the NTM problem i a EHWSN usig a path-costraied mobile sik. Due to some urealistic assumptios cosidered i [9] ad [10] ad i order to further improve the etwork throughput, the authors i [14] propose a coditio o the fixed distace travelled by sik at all give time slots. Based o the proposed coditio, they develop a olie cetralized algorithm which ot oly has low computatioal complexity but outperforms the previous works i term of etwork throughput. Motivated by the drawback of cosiderig fixed distace travelled by sik ad its costat speed i [14], i order to further improve the throughput, we propose differet scearios for NTM problem for which practical parameters are optimally utilized. Furthermore, due to the ifeasibility of assumptio o costat eergy harvestig i practical data collectio scearios [12], i cotrast to [14], we cosider a uiform eergy harvestig distributio i this work to statistically model the amout of harvested eergy by odes durig differet time itervals. PROPOSED WORK As wireless sesor etworks have special approach with battery power ad the coservatio of eergy. Use of eergy harvestig reduces the limitatios of battery. Due time varyig characteristic of sesor ode eergy chages every time ad will have differet power levels. Each ode has their idividual parameters such as rage, distace, locatio its trasmissio power level. So due to idepedet parameter of each sesor output must be as idepedet as possible, to avoid the covergece to the same maxima. So proposed algorithm estimates the idepedet compoets of the sesors its power level itesity by comparig the eergy with reliability. Network is said to cosistig of sesors with each havig a capability of reewig their eergy from surroudigs. A model is derived from the above assumptio ad stated below. A ode assumed to be havig storig eergy capacity of C ad eergy stored at give time iterval is P. With the eergy reewable model H, eergy at ay time P i (e) is defied usig (1), P i (e) = mi{p i 1 (e) + H i 1 (e) D i 1 (e), C(e)} (1) where, H i 1 (e) = harvested eergy, D i 1 (e) = depleted eergy, P i (e) With the assumptio as give (2) = stored eergy at each ode 0 P i (e) C(e) (2) It is well kow that sigals i wireless systems suffer from path loss, iterferece, shadowig ad other evirometal destructios. Sigal to Noise Ratio (SNR) determies the performace of the etwork. The system is reliable if all data packets reach to destiatio ode without ay drop. But, efficiecy of the etwork may go dow as system is more prioritized to esure reliability. Higher eergy ivestmet may yield improvemet i the reliability but itroduces uecessary eergy cosumptio. Takig reliability ad efficiecy as two odds, we itroduce multi speed trasmissio of mobisik that adds optimal usage of resource also providig reliability to the etwork. Give a ode i ad sik ode s, the trasmissio power ca be operated i multi-level which is set depedig the QoS of etwork i terms of reliability. It is stated before that, reliability icreases with icrease i power ad decreases if the distace betwee refereced odes is higher. Both combiedly defied usig (3), r i,j P i d α (3) i,j where, r i,j = reliability P i = Power cosumptio at i d = distace betwee ode i ad sik ode α = path loss factor Decisio makig is a importat parameter to decide trasmissio rate of a sesor ode. Variable trasmissio rate ca be set if certai coditios are met. Therefore, selectio of trasmissio rate will have decided by the amout of data a ode is holdig. Assumig, DT i is the amout of data ode i is about to sed, the DT i r i. Thus, reliability ca be related to amout of data beig set. 6656

3 Problem Defiitio: Network is defied to be havig eergy harvestig system H ad umber of time slots T assiged by mobisik ode per tour. Data collectio maximizatio is to allow the mobisik ode to collect maximum data from all odes withi the etwork withi the specified time slots. Reewal of eergy ad expese of eergy for trasmissio are the two costraits decides data collectio strategy. Ituitively, every sesor ode withi the etwork should trasmit data to mobisik withi the time slots allocated. However, the eergy repleishmet of each ode varies from other. Hece, it may lead to cosumptio of more time slots for few odes to trasmit data. Agai, eergy reewal rate will be slower tha trasmissio rate which is importat to desig the protocol. There are possibilities that multiple sesor ode may participate at same time to trasmit data ad compete each other resultig i collisio of packets. With all the above parameters, data collectio maximizatio problem ca be described below. Give, ode i N total umber of odes, umber of time slots τ, data collectio is to allocate the time slots such that i N. j N i (x i,j r i,j τ) will be maximum, with costraits give i equatios (4), (5), (6) x i,j {0, 1}, i N, j N i (4) x i,j = 0, i N, j N i (5) i=1 x i,j 1, i N, 1 j N i (6) j=1 P i,j x i,j τ P i (e), i N (7) τ = DT i τ u (8) where, x i,j = { 1, whe time slot t i is allocated to i 0, otherwise τ = umber of time slots τ u = sigle uit of time slot N i = eighbor ode set of ode i P i,j = power of ode i to trasmit data to ode j N i Proposed Algorithm: With the above mathematical model derived, the framework for the proposed algorithm preseted here. Network with N umber of odes withi the etwork ad harvestig eergy H total legth travelled by mobisik ode is termed as tours. Each tour will cover the etire etwork with predefied path by mobisik ode. I each tour, mobisik divides the total time ito time slots ad allocate time slots to all the odes withi the etwork. Allocatio of time slots is based o eergy repleishmet ad amout of data each ode carries. Higher the rate of eergy repleishmet, more umber of time slots will be allocated as it will have more eergy to ivest for data trasmissio. Fig shows the flow diagram of the proposed algorithm where mobisik ode will iitiate the tour ad divides the time slots. After the time slots are framed, mobisik will broadcast this iformatio to all odes withi the etwork. Every ode after receivig the packet will fid the slots ad tour patter of mobisik. Every data trasmissio results i decremet of eergy of a ode derived from the equatio (9) P i = P i j=1 P i, j τ (9) PERFORMANCE ANALYSIS Simulatio parameter Sl. No. Parameter Table 1: Simulatio Parameters Values 1 Network size (50 150) 2 Packet trasmissio rate 250 Kbps 3 Maximum trasmissio rage 4 Amout of eergy collected by EH 200 m mwh (suy day) mwh (cloudy day) 5 Solar pael size 10 mm 10 mm 6 Ambiet temperature Default time slot duratio 2 secod 6657

4 Remaiig Eergy: will be slight decrease will icrease of source odes. But, proposed algorithm shows better results compared to ICGS. Because, the tour trajectory will be iformed well before mobisik begis its travel. Hece, odes which are fall closer to the route path will be otified i prior ad esured that data will be reached to those odes by source ode ahead of time. This greatly improves the performace of the system i terms of deliverig the packets to mobisik. Throughput: Fig Remaiig Eergy of Source Nodes Fig shows the remaiig eergy of source odes withi the etwork. The results are compared with ICGS protocol. Participatio of odes i data trasmissio is proportioated with their eergy repleishmet as compared to ICGS where reewal eergy ad ode participatio i data trasmissio has ot related each other. If the source ode is ot able to get eough reewable source of eergy, the will just forward the data to eighbour odes who has more opportuity of gettig reewable eergy. Node which receive this data will take direct participatio with mobisik to deliver the data. Thus, source ode eergy will ot be expeded much if they are isufficiet to gai eergy from surroudigs. Hece, the result show better performace compared to ICGS protocol. Packet Delivery Ratio: Fig shows the results of packet delivery ratio with respect to source odes. Icrease i umber of source ode lead to allocatio of more time slots. But, the tour iterval is fixed, the there might be delay i deliverig the data. Hece, there Fig shows the results of throughput with respect to umber of odes i the etwork. It is already stated that, odes which have more capability of repleishig their eergy will participate i data trasmissio. Ad, it is proved from the proposed algorithm that, reliability is directly proportioal to expese of eergy. Therefore, throughput of the etwork gives satisfactory results compared to ICGS protocol. CONCLUSION I this paper, we have proposed a efficiet algorithm to maximize data collectio of eergy harvestig wireless sesor etwork usig mobile sik. Iitially, a problem is defied to uderstad the use case to icrease data collectio rate i a etwork. Problem is formulated further to derive a relatioship with reewable eergy ad amout of data to be trasferred. With all these parameters cosidered, it is prove that data collectio ca be maximized usig mobisik ode ad reewable eergy subject to more ivolvemet of high eergy ode i data trasmissio. The reults prove that data collectio rate improved upto 35% compared with ICGS protocol. Proposed algorithm is efficiet ad scalable where it does ot have depedecy o either locatio or volume of odes i the etwork. Further improvemet ca be doe by usig differet algorithm with differet routig protocols to optimize eergy ad data buffers for future work. It will improve the performace lifetime of a etwork with maximizig data collectio ad will make etwork reliable ad more efficiet. 6658

5 REFERENCES [1] M. Garey, ad D. Johso, Computers ad Itractability: A Guide to the Theory of NP- Completeess, Bell Laboratories Icorporated, [2] T. H. Corme, C. E. Leiserso, R. L. Rivest, ad C. Stei, Itroductio to Algorithms, MIT Press, 3rd Editio, [3] W. Liag, J. Luo, ad X. Xu. Prologig etwork lifetime via a cotrolled mobile sik i wireless sesor etworks. Proc. Of GLOBECOM, IEEE,2010. [4] M. D. Fracesco, S. K. Das, ad G. Aastasi, Data collectio i wireless sesor etworks with mobile elemets: a survey, ACM Tras. o Sesor Netw., [5] W. Liag ad J. Luo. Network lifetime maximizatio i sesor etworks with multiple mobile siks. Proc.OfLCN 11, IEEE,2011. [6] S. Gao, H.Zhag, ad S. K. Das, Efficiet data collectio i wireless sesor etworks with pathcostraied mobile siks,ieee Tras. O Mobile Comput., vol.10,2011. [7] R. S. Liu, K. W. Fa, Z. Zheg ad P. Siha. Perpetual ad fair data collectio for evirometal eergy harvestig sesoretworks, IEEE/ACM Tras. o Netw., vol. 19, [8] X. Re ad W. Liag, Delay-tolerat data gatherig i eergy harvestig sesor etworks with a mobile sik,, i Proc. Of IEEE GLOBECOM Coferece, [9] X. Re, ad W. Liag, The use of a mobile sik for quality data collectio i eergy harvestig sesor etworks, i Proc. Of IEEE Wireless Commu. ad Netw. Coferece (WCNC), [10] X. Re, W. Liag, ad W. Xu, Use of a mobile sik for maximizig data collectio i eergy harvestig sesor etworks, i Proc. of IEEE 42d It. Cof. o Parallel Processig, [11] L. He, J. Pa ad J. Xu, A progressive approach to reducig data collectio latecy i wireless sesor etworks with mobile elemets, IEEE Tras. Mobile Comput., vol. 12, [12] D. Guduz, K. Stamatiou, N. Michelusi, ad M. Zorzi, Desigig Itelliget Eergy Harvestig Commuicatio Systems, IEEE Commuicatios Magazie, vol. 52,2014. [13] J. Che, S. He, ad Y. Su, Rechargeable Sesor Networks: Techology, Theory, ad Applicatio, World Scietific Publishig,2014. [14] Mehrabi, ad K. Kim, Maximizig data collectio throughput o a path i eergy harvestig sesor etworks 6659