Delay Tolerance and Energy Saving in Wireless Sensor...

Research ArticleDelay Tolerance and Energy Saving in Wireless Sensor Networkswith a Mobile Base Station

Oday Jerew 1 and Nizar Al Bassam 2

1Asia Pacific International College Sydney Australia2Middle East College Muscat Oman

Correspondence should be addressed to Oday Jerew odayjerewgmailcom

Received 3 July 2018 Revised 30 October 2018 Accepted 17 December 2018 Published 12 February 2019

Academic Editor Donatella Darsena

Copyright copy 2019 Oday Jerew and Nizar Al Bassam This is an open access article distributed under the Creative CommonsAttribution License which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Recent research shows that significant energy saving can be achieved in wireless sensor networks by usingmobile devices Amobiledevice roams sensing fields and collects data from sensors through a short transmission range Multihop communication is usedto improve data gathering by reducing the tour length of the mobile device In this paper we study the trade-off between energysaving anddata gathering latency inwireless sensor networks In particular we examine the balance between the relayhop count andthe tour length of a mobile Base Station (BS) We propose two heuristic algorithms Adjacent Tree-Bounded Hop Algorithm (AT-BHA) andFarthestNode First-BoundedHopAlgorithm (FNF-BHA) to reduce energy consumption of sensor nodesThe proposedalgorithms select groups of Collection Trees (CTs) and a subset of Collection Location (CL) sensor nodes to buffer and forwarddata to the mobile BS when it arrives Each CL node receives sensing data from its CT nodes within bounded hop count Extensiveexperiments by simulation are conducted to evaluate the performance of the proposed algorithms against another heuristic Wedemonstrate that the proposed algorithms outperform the existing work with the mean of the length of mobile BS tour

1 Introduction

Data gathering in Wireless Sensor Networks (WSNs) is oneof the most frequent and fundamental operations whichrequires the sensor nodes to monitor the sensing field foras long as possible As sensor nodes have limited energyresources and are powered by small batteries energy con-sumption is a critical issue in the design of WSNs that affectthe energy consumption of sensor nodes and hence thenetwork lifetime

Recent research [1ndash5] shows that significant energy savingcan be achieved in wireless sensor networks by using mobiledevices capable of carrying data mechanically Mobile BaseStations (BSs) are proposed to reduce energy consumptionby allowing a mobile BS to roam a sensing field and gatherdata from sensor nodes through a short transmission rangeThe energy consumption of each sensor node is then reducedsince fewer relays are needed for the sensor node to relayits data packet to the BS As the speed of mobile BS is veryslow compared with the speed of data packets which travel in

multihop forwarding the increased latency of data gatheringwhen employing mobile BS presents a major performancebottleneck Consequently the time a mobile BS takes totour a large sensing field may not meet the stringent delayrequirements inherent in some mission-critical real-timeapplications Therefore planning the tour of mobile BS needsto be considered in order to achieve the delay requirements

Generally the mobility of BS can be classified into threetypes according to the mobility pattern of the entity uponwhich the BS is mounted as follows

(1) Random mobility this can be achieved for examplewhen the BS is mounted on humans and animalsIn this case the probability of the BS collecting allsensing data is low since the BS opportunisticallyvisits sensor nodes [6]

(2) Predictable mobility in this case the BS is mountedon an entity that moves on a fixed track or path thatcannot control its direction or speed butwhichmovesat a regular time for example a BS mounted on a bus

HindawiWireless Communications and Mobile ComputingVolume 2019 Article ID 3929876 12 pageshttpsdoiorg10115520193929876

2 Wireless Communications and Mobile Computing

or train Thus the sensor nodes can predict when theBS may move around to send their data [7 8]

(3) Controlled mobility when the BS is mounted on arobot or UAV plane then the direction and speedof the BS can be controlled Many algorithms areproposed to find the tour of the BS in order to achieverequirements such as maximised network lifetimeand data-gathering delay using single and multihoprelays [3 5 9]

Most literature studies the mobility of WSNs to bepredictable or controlled to achieve network performancerequirements (some literature focused on randommobility toimprove network lifetime) This differs from mobile ad hocnetworks in which nodes are assumed to move arbitrarilywhich degrades the network performance by link failures

In this paper we deal with the data gathering problemwith bounded delay time in sensor network using a controlledmobile BS by exploring a balance between the relay hopcount required for data relay and tour length of the mobile BSsuch that the energy consumption is minimized Specificallywe assume themobile BS selected a set of Collection Location(CL) nodes that temporarily cache and forward sensing dataof other nodesThe sensor nodes are grouped into CollectionTrees (CTs) rooted at CLs Each node has to forward its datato the CL within a certain number of relay hops The mobileBS traverses along a close tour and stops at each CL in thetour for data gathering

Our major contributions in this paper are as followsWe propose two efficient heuristic algorithms Adjacent Tree-BoundedHopAlgorithm (AT-BHA) and FarthestNode First-Bounded Hop Algorithm (FNF-BHA) The algorithms findthe tour of the mobile BS consisting of CLs which meet thefollowing criteria (i) the energy consumption among thesensor nodes is balanced in order to prolong nodes lifetime(ii) the total traversal time of the mobile BS on the touris minimized and (iii) the hop count between any sensornode and the CL is bounded by a given value The AT-BHAconstructs adjacent CTs by finding closest nodes to the lastconstructed CL The FNF-BHA uses the farthest node fromthe BS in the construction of CL Load balancing algorithm isproposed to balance the number of nodes in the CTs Finallythe performance of the proposed algorithms is evaluatedthrough experimental simulations The experimental resultsdemonstrate that the proposed algorithms are efficient

The remainder of this paper is organized as followsSection 2 reviews the related work on mobile data gatheringSection 3 presents two algorithms to solve data gatheringwith bounded hop count problem Section 4 evaluates theefficiency of the proposed algorithms through extensivesimulations Finally Section 5 concludes the paper

2 Related Work

There is an extensive range of literature studying the useof mobile BS for data gathering [4 5 9ndash15] Kansal et al[12] combine multihop forwarding with a mobile BS Theyperformed an experimental evaluation for a small sensornetwork assuming that a mobile BS moves back and forth

on a straight line (a fixed path) They employed a directeddiffusion approach to gathering sensed data from the sensornodes beyond the transmission range of the mobile BS

Some existing studies assume that sensor nodes can cachesensing data of other nodes in order to forward it to themobile BS when the BS moves around Gao and Zhang [11]and Xing et al [16] explored the delay requirement for datagathering assuming that the sensor nodes have an ability tocache sensing data of other nodes In [11] sensor nodes sendtheir data through multihop relays to the nodes (subsink)located within the direct-transmission range of the mobileBS The subsink nodes cache data and send it to the mobileBS when it comes within transmission range In [16] arendezvous-based data-gathering approach is proposed inwhich a subset of nodes is chosen as rendezvous points Therole of these points is to buffer and aggregate data originatingfrom sensor nodes When the mobile BS arrives within thetransmission range of rendezvous points the data will beforwarded to the mobile BS Kaswanet et al [17] proposedpath plan algorithm for data collection by selecting set ofrendezvous points The algorithm minimize the tour lengthby reducing the number of rendezvous points by selecting theposition closest to the location of the BS and allows the BS tocollect data frommaximum number of sensors via single hopcommunication This algorithm does not consider a boundedhop count for data collection

Grouping sensor nodes in the sensing field into clustersand collecting data from cluster heads are proposed in [13ndash15]Ma and Yang [13] proposed a heuristic for finding routingpaths for the mobile BS which assumes that the moving pathof the mobile BS consists of a series of line segments Sensornodes closest to each line segment are selected as clusterheads A specified configuration is applied where the mobileBS starts data gathering from the left side of the path movestowards the right side and then comes back to the left sideagain This scheme maximizes network lifetime However itmay cause packet losses at each cluster head if the path lengthto the cluster head is too long since the time required for thecluster head to send the data of cluster nodes to the BS is notconsidered in cluster forming Saad et al [14] proposed pathplanning algorithm for mobile BS that visit a set of clusterheads Nodes are randomly grouped to form clusters and thenode with the largest residual energy is selected as the clusterhead The clusters are merged ensuring that the number ofhops from the cluster head to any cluster node is no morethan two hops The cluster head gathers and buffers clustersensing data The BS visits the cluster heads to gather datathrough single-hop communicationHowever thiswork doesnot consider data gathering delay for the mobile BS Zhang etal [15] extended the work in [16] and regarded cluster headsas the rendezvous point They considered event based datagathering and proposed reduction in data gathering delayby only allowing the mobile BS to visit cluster heads thatgenerate new data

Bounded hop count for data gathering is studied in [45 18 19] Zhao and Yang [5] proposed SPT-DGA and BRH-MDG data gathering algorithms that select set of CL for themobile BS to visit for data gathering Each CL aggregatesthe local data from its affiliated sensors within a certain

Wireless Communications and Mobile Computing 3

number of relay hops However the number of nodes inthe constructed CT is small Consequently many CTs areconstructed for data gathering Thus the length of the BStour increased Chen et al [19] analysed and improved theBRH-MDGalgorithmbyminimizing the degree of the CTs inorder to prolong the network lifetime Xu et al [4] proposedheuristic algorithm to find the tour for a mobile BS consistsof CL nodes However the algorithm does not consider theeffect of some critical sensor nodes on the separation ofother nodes These critical nodes affect the location of CLsand hence the length of the mobile BS The hop count andhop distance are used to plan the route of the mobile BSin [18] In this work set of rendezvous points are selectedsuch that the mobile BS can collect data from the sensornodes within a permissible delay with minimum possiblehop distance and hop counts However this work does notconsider minimizing the BS tour length

Dividing sensing field into subareas is proposed in [2021] Chen et al [20] introduced algorithm to improve thenetwork lifetime considering data transmission delay andnumber of hops The algorithm divided the sensing fieldinto grids the mobile sink has to stay at the grid centerfor data collection However this work assumes the sensornodes send data to the BS only when the hop count betweenthe sensor and the BS is at or below a specific hop countOtherwise the sensor enters a sleeping state In addition thehop count is estimated using the distance between the sensornodes and the BS which is not accurate at low node density(node degree less than six) A delay bound path algorithm isproposed in [21]The algorithm divided the sensing field intohexagonal cells whose centers are considered as the positionsof rendezvous point The algorithm minimizes the numberof rendezvous point by selecting the location that coversas many number of sensor nodes and at minimal distancefrom the center Data gathering within bounded hop countis also studied in [22] The authors proposed reducing thenumber of collection points by selection of the CL within theconvergence area that overlaps maximum number of sensornodes However the work in [21 22] does not consider theeffect of isolated sensor nodes on the tour length of the BS

Other literature involved sensor node residual energy andbounded hop count in the construction of the data gatheringtrees [2 23] Zhu et al [23] suggested assigning weight foreach node based on nodes residual energy distance to the BSand number of hops The weight is used to select rendezvouspoints for data collection Jerew and Al Bassam [2] examinedthe sensor nodes residual energy to select a set of cluster headswith high residual energy and bounded hop count Sensornode sends its data within bounded hop count to the clusterhead and the BS visits all the cluster heads for data gatheringChang and Shen [24] introduced a cluster routing structureto decide the routing path based on the location of BS thedistances between the sensor nodes and the residual energyof each sensor node The closest node to the BS with residualenergy more than the average network residual energy isselected as cluster head The proposed algorithm reducedthe energy consumption and extended the sensor lifetimeby balancing the network load However this work did notconsider data gathering within a bounded hop count

21 Preliminaries and Problem Weconsider a wireless sensornetwork 119866 = (119881 119864) consisting of 119899 = |119881| stationary sensorsand 119864 is the set of links There is a link between two sensorsor a sensor and the BS if they are within the transmissionrange of each other For the sake of simplicity we assumethe transmission range of each sensor node and the BS arefixed and identical and all sensor nodes have identical initialenergy The storage of a sensor node is limited so that itcannot buffer a large volumeof data Sensor nodes are denselydeployed in the sensing region Accordingly the number ofhops in a path is approximately proportional to the distancebetween the nodes The BS moves with constant velocityThus there is sufficient time to establish communicationand send one or more data packets during the time theBS takes to travel across the transmission range of a sensornodeWe further assume themobile BS replenishes its energyperiodically so there is no energy concern with themobile BSFinally sensor nodes and the mobile BS are assumed to knowtheir own physical locations via GPS or a location service inthe network

Problem Given a network with a mobile BS and assum-ing that the number of hops for data gathering between theBS and sensor nodes is bounded by ℎ the problem is tofind the shortest tour for the mobile BS such that the energyconsumption among the sensor nodes is balanced

3 Proposed Heuristic Algorithms

Due to the NP-hardness of this problem [5] a heuristicalgorithm is proposed in this paper In order to find theoptimal CLs among sensors relay routing paths and the tourof the mobile BS should jointly be considered Based onthese observations we deal with the tour lengthminimizationproblem by devising scalable heuristic algorithms In thissection we proposed two algorithms the Adjacent Tree-Bounded Hop Algorithm (AT-BHA) and the Farthest NodeFirst-Bounded Hop Algorithm (FNF-BHA) Table 1 showsthe main symbols used in the proposed algorithms

31 Adjacent Tree-Bounded Hop Algorithm (AT-BHA) Thebasic idea of AT-BHA algorithm is to find a set of CLs in thenetwork that the BS can visit to collect sensing data of all thesensors Visiting CLs in specific order determines the lengthof the tour of the mobile BS The value of ℎ determines thedelay on data delivery and affects the number of sensor nodescontained in each tree and hence the number of CLs Forexample if ℎ = 1 the number of CLs is equal to the numberof sensor nodes and the BS tour is the longest tour as the BShas to visit each node for one hop data collection

In any network the node degree (number of neighbornodes) significantly affects the network connectivity Asthe node degree 119889 lt 6 there is a high probability ofpartitioned networks [25]The first task of AT-BHA as shownin Algorithm 1 is to construct Minimum Spanning Trees(MSTs) that cover all sensors in the network Each sensornode associated in a single MST referred to as 119879119875 Thusthe AT-BHA can be applied to connected and disconnectednetworks


Input A sensor network 119866(119881 119864) hop cout ℎ 119903 120573Output Set of CL nodes V119862119871 119861119879119862(1)119873119866 larr997888 119881 lowastset of nodes in the network lowast(2) 119896 = 0 (3) 119895 = 0 (4) while 119873119866 is not empty do(5) Find the closest vertex V119888 to the BS position from119873119866 (6) Find an MST 119879119875 that covers vertices from119873119866 (7) 119873119875 larr997888 119879119875 lowast set of nodes in partition tree 119879119875 lowast (8) while 119873119875 is not empty do(9) Find the farthest leave vertex V119891 in 119879119875 (10) Find the routing path 119877(V119891 V119888) (11) if ℎ ge |119877(V119891 V119888)| then(12) V119862119871 larr997888 V119888 (13) else(14) V119862119871 larr997888 vertex from 119877(V119891 V119888) at ℎ hops from V119891 (15) 119895 = 119895 + 1 (16) Find an MST 119879119862119895ℎ(V119862119871 119873119875) (17) 119873119875 larr997888 119873119875 minus 119879119862119895ℎ (18) if 119873119875 is empty then(19) break(20) Find 119899119891 node that with119872119886119909119889(119873119875 V119888)119889(119873119875 V119862119871) (21) Find119873119888119903 set of nodes with 119889(119873119875 119899119891) le ℎ119903120573 lowast nodes within a circle of ℎ119903120573 radius lowast (22) Find the closest vertex V119899 to V119888 from119873119888119903 (23) 119895 = 119895 + 1 (24) V119862119871 larr997888 V119899 lowast set V119899 as a CL lowast (25) Find an MST 119879119862119895ℎ(V119862119871 119873119875) (26) 119873119875 larr997888 119873119875 minus 119879119862119895ℎ (27) (119879119862119873119875)=Update CT(V119862119871119895119873119875119879119862119879119875) (28) 119873119866 larr997888 119873119866 minus 119873119875(29) [119861119879119862 V119862119871]=Call LBA(V119862119871)

Algorithm 1 The Adjacent Tree-Bounded Hop Algorithm (AT-BHA)

Table 1 Definitions of the main symbols used throughout thealgorithms

Symbol Description119881 Set of network nodes120573 Between -1 and 1 to calculate hop progress120591 Threshold value to decide tree updateℎ Number of hops119879119875 Set of nodes in a partition tree

V119888Root node of the MST that is closest to the location

of BSV119891 Farthest node from the BS119877(V119891 V119888) Set of nodes in the routing path between V119891 and V119888V119862119871 Set of the collection location nodes

119879119862119895ℎSet of nodes contained in the partial 119895th tree within ℎ

hops away from the root119873119875 Set of the nodes at network partition tree119889(V1 V2) Euclidean distance between vertices V1 and V2119899119891 Farthest node to the BS and closest to the 119879119862119895ℎ tree119903 Transmission range of a sensor node119873119903 Set of nodes to be removed from a tree

The root of the MST is selected at sensor node V119888 that isthe closest to the location of the BS We set the root close tothe BS in order to select CLs as close as possible to each otherand close to the BS initial position

The next task of AT-BHA is to find the farthest node V119891from the BS position in the partition tree The routing pathbetween the farthest node and root of 119879119875 is referred to as119877(V119891 V119888) and it is used to find the candidate CL node Weassume the set of nodes in 119877 is sorted in decreasing orderaccording to the number of hops to V119888 In the case of thenumber of nodes in the routing path being less than ℎ thenthe root of the 119879119875 tree is selected as V119862119871 While the node at ℎhop away from the V119891 is selected as V119862119871 if the route length ishigher than ℎ

The Breath-First-Search (BFS) algorithm with boundedhop count is used to find theMinimum Spanning Tree (MST)rooted at the CL A partial BFS tree is constructed hop by hopand the expansion continues until all nodes within ℎ hops areexplored The set of sensors contained in the partial 119895th treeis referred to as 119879119862119895ℎ The set of the nodes covered in the 119895thCT is removed from the set of partition tree nodes119873119875

In this algorithm we propose to find the next CT adjacentto the last constructed tree in order to avoid partitioned node


Input V119862119871 119895119873119875 119879119862 119879119875 120591Output 119879119862119873119875(1) if V119862119871 is the first time for check then(2) 119873119862119871 larr997888 119879119862119895ℎ (3) if |119873119862119871|119889(V119862119871 V119888) le 120591 then(4) Find the routing path 119877(V119862119871 V119888) using partition tree 119879119875 (5) if ℎ ge |119877(V119862119871 V119888)| then(6) 119873119875 larr997888 119873119875 cup 119877(V119862119871 V119888) (7) V119862119871 larr997888 V119888 (8) else(9) Find vertex Vℎ from 119877(V119862119871 V119888) at ℎ hops from V119862119871 (10) 119873119875 larr997888 119873119875 cup 119877(V119862119871 Vℎ) (11) V119862119871 larr997888 Vℎ (12)(13) Find an MST 119879119862119895ℎ(V119862119871 119873119875) lowast Update the j-th treelowast (14) 119873119875 larr997888 119873119875 minus 119879119862119895ℎ (15) Return(119879119862119873119875)

Algorithm 2 Update CT

between the CTs The partitioned node increases the tourlength since the BS has to visit the node for data collection Tofind an adjacent tree to the 119895119905ℎ tree the algorithm calculatesthe ratio 120572 = 119889(119873119875 V119888)119889(119873119875 V119862119871) The 119889(V1 V2) representsEuclidean distance between vertices V1 and V2The node from119873119875 with maximum 120572 is selected and referred to as 119899119891 Thisnode represents the farthest node to BS and closest to the119879119862119895ℎ The closest node to the BS that is within the distanceℎ119903120573 from 119899119891 is selected as the second V119862119871 where 119903 is thesensor transmission range and 120573 between minus1 and 1The value119903120573 represents the hop progress and its ranges from minus119903 to119903 depend on the network node density [25] New MST isconstructed rooted at V119862119871 to find set of nodes within ℎ hops

The number of nodes within each tree and the positionof the CL nodes have a signification effect on the tour lengthWhen the CT consists of many nodes a less number of CLsare required and the tour length decreases In addition thetour length is also decreased as the selected CLs are closeto the BS To improve the tree size and location of V119862119871 theUpdate CT algorithm is proposed in Algorithm 2

In the Update CT algorithm the ratio of the number oftree nodes to the distance between the V119862119871 and V119888 is usedas a measure of tree usefulness A threshold value 120591 isused to decide the updating of the CT The range of 120591 is(1119889(119871119861119878 119873119865) 119899119889(119871119861119878 V119861119878)) where 119899 is the network nodesand 119871119861119878 V119861119878 and119873119865 are respectively the position of BS theclosest node to the BS and the farthest node to the BS

In order to avoid infinite looping the algorithm firstchecks whether the CT is previously updated since in somecases it is impossible to update the tree because of limitednumber of nodes In the case the CT needs to update therouting path between V119862119871 and V119888 is used to select a new CLnode119873119875 is updated by including the set of CT nodes and theset of routing nodes to the selected V119862119871 Finally the updatedCT is constructed and returns the updated CT and119873119875 to theAT-BHA algorithm The AT-BHA algorithm continues untilall the network nodes are associated with CTs

32 Farthest Node First-Bounded Hop Algorithm (FNF-BHA)This algorithm constructs CTs to include the farthest nodesto the BSThe FNF-BHA starts similar to AT-BHA by findingthe nodes closest to the position of the BS and constructionpartitions trees It then finds the route of data packets to V119891in the current partition tree to determine the CL node Thenodes involved in the data relay between V119891 and the collectionnode are critical for data gathering from V119891 In the case ofany of the relay node participating in another CT the BShas to visit V119891 for data collection increasing the tour lengthTherefore in this case the algorithm first identifies and thenremoves all the relay nodes from other CTs by calling theNode Removal Algorithm (NRA) as shown in Algorithm 3The set of removed nodes is used in the construction of newCTs at the next iteration

The algorithm tests whether the CL node is alreadyselected as CL of another tree In this rare case the set of CTnodes are merged with that tree and new CT is constructedThenew tree includes all the nodes of previousCT in additionto V119891 since all the relay nodes are available in the set ofnodes whereas if the CL node is not selected a new CTis constructed with a new tree index to include V119891 Thealgorithm is terminated when all the partitioned tree nodesare covered

We now describe NRA algorithm in more detail Thepseudocode is given in Algorithm 4 In order to remove theset of nodes in 119873119903 from the CTs the tree index of each nodeneeds to be found first The set of nodes that need to beremoved fromeachCTare identified and removedA newCTis reconstructed after node removal In some cases some ofthe tree nodes 119873119900 are partitioned since the removed nodesprovide a relay path for them 119873119900 may be covered by otherCTs Therefore all the CTs are reconstructed considering 119873119900If a node is associated with CT the node is removed from119873119900The algorithmcontinues until all the nodes in119873119903 are removedfrom the CTs


Input A sensor network 119866(119881 119864) hop cout ℎOutput Set of CL nodes V119862119871 119861119879119862(1)119873119866 larr997888 119881 lowastset of nodes in the networklowast (2) while 119873119866 is not empty do(3) Find the closest vertex V119888 to the BS position from119873119866 (4) Find an MST 119879119901 that covers vertices from119873119866 (5) 119873119875 larr997888 119879119875 set of nodes in partition tree Tp (6) while 119873119875 is not empty do(7) Flaglarr997888 false lowastno node need to remove from any tree (8) Find the farthest leave vertex V119891 to V119888 on 119879119901 (9) Find the routing path 119877(V119891 V119888) (10) if ℎ ge |119877(V119891 V119888)| then(11) V119862119871 larr997888 V119888 (12) else(13) V119862119871 larr997888 vertex from 119877(V119891 V119888) at ℎ hops from V119891 (14) 119873119903 larr997888 set of nodes in 119877(V119891 V119888119871) not in119873119875 (15) if 119873119903 is not empty then(16) Flaglarr997888 true (17) 119873119903119898 = NRA(119873119903 119873119875) lowast call Node Removal Algorithm lowast

119873119875 larr997888 119873119875 cup 119873119903119898 (18) if Flag= false then(19) if V119862119871 selected as 119862119871 then(20) 119895 =index of tree rooted on VC119871 (21) 119873119875 larr997888 119873119875 cup 119879119862119895ℎ (22) else(23) select V119862119871 as Collection Location (24) 119895 = 119895 + 1 (25)(26) Find an MST 119879119862119895ℎ(V119862119871 119873119875) (27) 119873119875 larr997888 119873119875 minus 119879119862119895ℎ (28) 119873119866 larr997888 119873119866 minus 119873119875 (29) [V119862119871119861119879119862]=Call LBA(V119862119871)

Algorithm 3 The Farthest Node First-Bounded Hop Algorithm (FNF-BHA)

Input A sensor network 119866(119881 119864) hop cout ℎ119873119903119873119875Output 119873119903 119879119862(1) 119868119883 larr997888 set of tree index of all CTs (2) 119868119883119903119905 larr997888 set of tree index of all119873119903 nodes (3)119873119903119898 larr997888 120601 (4)119873119900 larr997888 120601 (5) for each 119895 in 119868119883119903119905 do(6) 119873119903119905 larr997888 119873119903 cap 119879119862119895ℎlowast Set of nodes to remove from the j-th tree lowast (7) 119873119899119905 larr997888 119879119862119895ℎ minus 119873119903119905 lowast Set of node of the j-th tree after node removel lowast (8) get V119862119871 from tree 119895 (9) Find an MST 119879119862119895ℎ(V119862119871 119873119899119905) (10) 119873119875 larr997888 119873119899119905 minus 119879119862119895ℎ lowast Set of partitioned nodes lowast (11) 119873119900 larr997888 119873119900 cup 119873119875 (12) if 119873119900is not empty then(13) for each tree index 119896 in 119868119883 do(14) get V119862119871 for tree 119896 (15) 119899119905 larr997888 119879119862119896ℎ (16) 119899119904119890119905 larr997888 119879119862119896ℎ cup 119873119900 (17) Find an MST 119879119862119896ℎ(V119862119871 119899119904119890119905) (18) 119873119889119894119891 larr997888 119879119862119896ℎ minus 119899119905 (19) 119873119900 larr997888 119873119900 minus 119873119889119894119891 (20) 119873119903119898 larr997888 119873119903119898 cup 119873119903119905 cup 119873119900 (21) Return(119873119903119898)

Algorithm 4 Node Removal Algorithm (NRA)


Input A sensor network 119866(119881 119864) hop cout ℎ tree index 119895 119879119862Output Set of CL nodes 119862119871 119861119879119862(1) [119861119879119862119875119873]=Call CTBA(119866 119879119862) (2) while 119875119873 is not empty do(3) 119879119862 = 119861119879119862 copy all the CTs to 119879119862 (4) 119895 larr997888 number of trees in 119879119862 (5) Find the farthest leave vertex V119891 from 119875119873 using 119879119875 (6) Find the routing path 119877(V119891 V119888) (7) if ℎ ge |119877(V119891 V119888)| then(8) V119862119871 larr997888 V119888 (9) 119877119899 larr997888 |119877(V119891 V119888)| (10) else(11) V119862119871 larr997888 vertex from 119877(V119891 V119888) at ℎ hops from V119891 (12) 119877119899 larr997888 nodes at ℎ hops from 119881119891 (13) 119875119873 larr997888 119875119873 cup 119877119899 (14) 119895 = 119895 + 1 (15) Find an MST 119879119862119895ℎ(V119862119871 119875119873) (16) 119875119873 larr997888 119875119873 minus 119879119862119895ℎ (17) [119861119879119862119875119873]=Call CTBA(119866 119879119862) (18) Return(119862119871 119861119879119862)

Algorithm 5 Load balanced algorithm (LBA)

Input A sensor network 119866(119881 119864) Set of CTs 119879119862119895ℎ hop count ℎOutput set of balanced CTs 119861119879119862119895ℎ Remain Nodes 119875119873(1) 119870 larr997888 number of CL nodes (2)119873119866 larr997888 119881 lowast set of nodes in the network lowast (3)119873119866119861 larr997888 119873119866 (4) 119861119879119862119895ℎ119888119895 = 1 119870 ℎ119888 = 1 ℎ larr997888 120601 (5) Sort 119879119862119895 trees in increasing order according to the number of nodes in each tree (6)119873119866119861 = 119873119866 minus V119862119871 remove CL nodes from testing nodes (7) for hc= 1 to h do(8) for each V119862119871 do(9) 119895 = 119895 + 1 (10) if ℎ119888 gt 1 then(11) 119873119866119861 = 119873119866119861 cup 119861119879119862119895ℎ119888minus1 (12) Find an MST 119861119879119862119895ℎ119888(V119862119871 119873119866119861) (13) 119873119866119861 = 119873119866119861 minus 119861119879119862119895ℎ119888 lowast remove tree nodes from testing nodes lowast (14) 119875119873 = 119873119866 minus 119873119866119861 (15) Return(119861119879119862119875119873)

Algorithm 6 CT balanced algorithm (CTBA)

33 Load Balancing and Further Improvement Theproposedalgorithms AT-BHA and FNF-BHA can be further improvedby balancing the number of nodes in the CTs The proposedalgorithmsfind the set of CLnodes to collect data from sensornodes within ℎ hops However there are large differencesbetween the number of nodes in the CTs In particular thefirst constructed tree covers the maximum number of nodeswhile the last tree covers only the remaining set of nodes

Thus we proposed the load balanced algorithm (LBA)with the pseudocode given in Algorithm 5 First the algo-rithm balanced the number of nodes for each CT using theCT balanced algorithm (CTBA) shown in Algorithm 6 TheCTBA first sorts the CTs in increasing order according to

the number of nodes in order to start constructing the CTwith theminimumnumber of nodesThe algorithmgraduallyconstructs all the CTs by finding the nodes with one hop fromthe CL nodesThen the number of nodes in the CTs increaseswith the hop count In some cases some nodes does notassociate to any tree since they are partitioned as the relayingnodes are associated to other CTs Finally the LBA returnsthe balanced CTs (BTC) and partitioned nodes (PN) Whenthere is no PN the algorithm finalizes the BTC and no furtherprocess is required However a new set of V119862119871 needs to beselected from the PN for data collection This is achieved byfinding the farthest node V119891 to the BS from PN The routingpath between V119891 and closest node to the BS is used to find


the V119862119871 at ℎ hops A new CT is constructed and rooted at V119862119871to cover PN nodes Finally the algorithm rebalanced the loadby calling the LBA The LBA is terminated when there is nopartition node in the PN

4 Performance Evaluation

In this section we evaluate the performance of the proposedalgorithms through simulations We assume that sensornodes in the network are randomly deployed with uniformdistribution in a 400 119898 times 400 119898 square sensing field Eachsensor node has a transmission range of 119903 = 25 119898 Wefirst vary the hop count to evaluate the performance of thealgorithms We also vary the number of nodes in the networkto emulate the change in the node degree The node degree isused as a metric of node density We compare our algorithmswith the SPT-DGA algorithm proposed in [5] For the AT-BHA we set 120573 = 05 and 120591 = 01 For each instance ofdeployment the network performancemetrics are calculatedand the result is the average over 500 instances for each nodedegree and hop count

We adopt the nearest neighbor algorithm [26] in oursimulation for the travelling salesman problem to determinethe tour of the mobile BS In order to calculate the energyconsumption of CL in transmitting CT sensors data to theBS on wireless communication per time unit we adopt theenergy model [2] 119864119862119862119871(119862119871) = 119903119886 sdot 119871119877 sdot (119873119879119862119871 + 1) sdot 119864119905where 119864119862119862119871(119862119871) is the energy consumption of CL 119903119886 is datageneration rate 119871119877 is the length of single data reading and119873119879119862119871 is the number of sensors the CL node has to forwardtheir data to the BS 119864119905 is the amount of power consumedby transmitting a unit-length of data 119864119905 can be representedby 119864119905 = 1205721 + 1205722119903

120574 where 1205721 is a constant that representsthe energy consumption to run the transmitter circuitrywhich is negligibly small 1205722 is a constant that represents thetransmitter amplifier and 119903 is the transmission range Theexponent 120574 is determined by the field measurements whichis typically a constant between 2 and 4 In this simulation weassume 119903119886 = 1 119870119887119901119904 119871119877 = 1119870119887 1205721 = 0 1205722 = 150 1198991198691198871198941199051198982and 120574 = 2

41 Varying Hop Count We first study the effect of changingthe hop count on the network performance Figure 1 plotsthe performance of our algorithms and the SPT-DGA interms of tour length and hop count ℎ When ℎ is smallthe number of nodes within each CT is small and thus thenumber of CLs increased which increases the length of themobile BS The result also shows that the tour length inthe FNF-BHA algorithm is the shortest since the FNF-BHAalgorithm considered the farthest nodes

The maximum number of nodes in the CTs with thenumber of hops is shown in Figure 2 The result shows thatas the hop count increases the number of nodes in the CTsincreasedThismakes sense as increasing the number of hopsadded extra nodes to the CTs The result also shows thatthe number of nodes in the CTs constructed by SPT-DGAalgorithm is always less than that in our algorithms

We also study the number of CLs with the hop count asshown in Figure 3 The result shows that the numbers of CLs

Tour Length with Hop Count

0

500

1000

1500

2000

2500

3000

3500

4000

4500

Tour

Len

gth

m

3 4 5 62Hop count h

SPTminusDGAFNFminusBHAATminusBHA

Figure 1 The tour length of the mobile BS with the hop count forthe SPT-DGA FNF-BHA and AT-BHA algorithms for 119889 = 12

Max Number of Nodes with Hop Count

20

40

60

80

100

120

140

160

180M

ax N

umbe

r of N

odes

3 4 5 62Hop count h


Figure 2 The maximum number of nodes in the CT with the hopcount for the SPT-DGA FNF-BHAandAT-BHAalgorithms for119889 =12

in our algorithms are significantly less than that in SPT-DGAthis is because the proposed algorithms try to construct CTswith maximum number of nodes by finding routes to thepartition nodes and adjacent CTs

The CL consumes more energy than any other networknodes since it has to forward the sensing data of the CT to theBS Figure 4 illustrates the maximum energy consumptionof CL for forwarding sensing data of a single sensor nodeby dividing the total energy consumption of the CL to thenumber of sensors in the CTThe result shows that the energyconsumption decreases with the hop count as the number ofnodes in the CT increased with hop count The result alsoshows that the energy consumption of the CL at FNF-BHCalgorithm is lower that other algorithms since the FNF-BHC


No of Collection Locations with Hop Count


3 4 5 62Hop count h

0

20

40

60

80

100

120

140

No

of C

olle

ctio

n Lo

catio

ns

Figure 3The number of CLs with the hop count for the SPT-DGAFNF-BHA and AT-BHA algorithms for 119889 = 12

Max Energy Consumption of CL Per Sensor with Hop Count

SPT_DGAFNFminusBHCATminusBHC

3 4 5 62Hop count h

94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J

Figure 4Themaximumenergy consumption of theCL sensor nodefor forwarding sensing data of single sensor nodewith the hop countfor the SPT-DGA FNF-BHA and AT-BHA algorithms for 119889 = 12

algorithm constructed CTs with the maximum number ofnodes

The maximum distance between the CL and the farthestsensor nodes is examined in Figure 5 The CLs selected bythe AT-BHA are the closest to the sensor nodes since the AT-BHAalgorithmconstructed adjacent trees using hop progress120573119903 rather than dealing with the farthest nodes

To better understand the results we give examples inFigures 6 7 and 8 where 245 sensors are scattered over200 119898 times 200 119898 field (119889 = 12) with the initial position ofdata BS located at the center of the area The ℎ is set to 4

Max Distance with Hop Count


3 4 5 62Hop count h

0

20

40

60

80

100

120

140

Max

Dist

ance

m

Figure 5 The maximum distance between the CL and the farthestsensor node with the hop count for the SPT-DGA FNF-BHA andAT-BHA algorithms for 119889 = 12

0 50 100 150 200

Collection Locations and Collection Trees for SPTminusDGA

0

50

100

150

200

Figure 6 An example illustrates the CLs and CTs constructed usingSPT-DGA for a network of 244 sensors The tour length is equal to5992mwith 15 CL nodes ℎ = 4 and 119889 = 12

which means that it is required for each sensor to forwardits data to the CL within four hops The figures also show theconstruction of CTs rooted at CLs for each algorithm

The SPT-DGA constructs 15 routing trees for data collec-tion as shown in Figure 6 The CLs are selected very close tothe center of the sensing area in order to minimize the tourlength However for the same sensors positions the AT-BHAand FNF-BHAconstructed 6 and 5 routing trees respectivelyThe position of the CLs selected in the FNF-BHA is closer tothe center of the sensing field than that in AT-BHA

42 Varying Node Degree We then vary the node degreein the network by varying the number of network nodesaccording to 119899 = 119889 lowast 119860(120587 lowast 1199032) where 119860 is the networkarea and 119903 is the node transmission range


0 50 100 150 200

Collection Locations and Collection Trees for ATminusBHA

0

50

100

150

200

Figure 7 An example illustrates the CLs and CTs constructed usingAT-BHA for a network of 244 sensors The tour length is equal to4646mwith 6 CL nodes ℎ = 4 and 119889 = 12

0 50 100 150 200

Collection Locations and Collection Trees for FNFminusBHA

0

50

100

150

200

Figure 8 An example illustrates the CLs and CTs constructed usingFNF-BHA for a network of 244 sensors The tour length is equal to3605mwith 5 CL nodes ℎ = 4 and 119889 = 12

To study the effect of the node degree on the tourlength the tour of the mobile BS is calculated using thenearest neighbor algorithm (as shown in Figure 9 for ouralgorithms and the SPT-DGA algorithm) It can be seen thatunder different 119889 the tour length for AT-BHC and FNF-BHA is shorter than that for the SPT-DGA algorithm Thisis because the FNF-BHC selects the minimum number ofCLs considering the farthest nodes to the position of the BSwhile the SPT-DGA selects more CLs which increases thelength of mobile BS tour The result also shows that the tourlength slightly increases with the node degree in SPT-DGAHowever increasing the node degree reduced the tour lengthfor the AT-BHA and FNF-BHA since that improves networkconnectivity and finding routing paths to the farthest nodesThis reduces the number of CLs as shown in Figure 10

Figure 11 illustrates the maximum number of nodes in theCTs with the node degree as a function of ℎ The result showsthat the number of nodes increaseswith the node degree sincethat increases the number of neighbors

0

500

1000

1500

2000

2500

Tour

Len

gth

m

Tour Length with Node Degree


9 10 11 128Node Degree d

Figure 9The tour length of the mobile BS with the node degree forthe SPT-DGA AT-BHA and FNF-BHA algorithms for ℎ = 4

No of Collection Locations with Node Degree

0

10

20

30

40

50

60

70

No

of C

olle

ctio

n Lo

catio

ns



Figure 10 The number of CLs with the node degree for the SPT-DGA AT-BHA and FNF-BHA algorithms for ℎ = 4

Figure 12 shows themaximum energy consumption of theCL for forwarding sensing data of a single sensor node tothe BSThe result shows that the energy consumption slightlydecreases with the node degree In addition the CL at FNF-BHA consumes less energy than other algorithms since thenumber of nodes at the CT of the FNF-BHA is more thanthe number of nodes at the CTs of SPT-DGA and AT-BHA asshown in Figure 11

We study the distance the packet travels to reach the BSby calculating the distance between the sensor nodesThedis-tance between the farthest sensor node and the CL is shownin Figure 13 The result shows that the maximum distance ishigher in SPT-DGA and FNF-BHA than in AT-BHA Thisis because the expansion of CTs in AT-BHA depends on the


Max Number of Nodes with Node Degree

0

20

40

60

80

100

120

Max

Num

ber o

f Nod

es



Figure 11Themaximumnumber of nodes with the node degree forthe SPT-DGA AT-BHA and FNF-BHA algorithms for ℎ = 4

Max Energy Consumption of CL Per Sensor with Node Degree



94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J

Figure 12 The maximum energy consumption of the CL sensornode for forwarding sensingdata of single sensor nodewith the nodedegree for the SPT-DGA AT-BHA and FNF-BHA algorithms forℎ = 4

settings of 120573 and hop count while it depends on hop countonly for the SPT-DGA and FNF-BHA algorithms

Finally we study the effect of load balanced algorithm onthe distribution of the number of sensor nodes in the CTsFigure 14 shows the histogram of the number of nodes inthe CTs before and after the execution of the load balancedalgorithm The result shows that there is high occurrence oftrees with the number of nodes less than 20 nodes for theunbalanced FNF-BHAHowever the number of nodes is verywell distributed for the balanced FNF-BHA

0

10

20

30

40

50

60

70

80

90

Max

Dist

ance

m

Max Distance with Node Degree



Figure 13Themaximumdistance between theCL and sensor nodeswith the node degree for the SPT-DGA AT-BHA and FNF-BHAalgorithms for ℎ = 4

Balanced and Unbalanced Collection Trees

Unbalanced FNFminusBHABalanced FNFminusBHA

0

50

100

150

200

250

300

350

400

450

Num

ber o

f Occ

urre

nce

20 40 60 80 100 1200Number of Nodes in the Collection Trees

Figure 14 Comparison between the number of nodes at balancedand unbalancedCTs for FNF-BHAalgorithmwith ℎ = 4 and 119889 = 10

5 Conclusion

In this paper we dealt with the problem of data gathering inthemobile BS environmentWe studied the trade-off betweenthe relay hop count of sensor nodes and the tour length ofthe mobile BS Two heuristic algorithms AT-BHA and FNF-BHA are proposed towards finding the shortest tour lengthfor the mobile BS subject to bounded hop count The AT-BHA selects adjacent CTs during tree construction to avoidpartitioned nodes that increased the tour length The FNF-BHA uses the farthest nodes from the BS in the constructionof the CTs Extensive simulations have been carried out tovalidate the efficiency of the proposed algorithms againstSPT-DGA The experimental results demonstrated that the


proposed algorithms outperform the SPT-DGA significantlyin terms of minimizing the BS tour length

Data Availability

The data used in this study are generated via the simulationof the proposed algorithms The proposed algorithms areprovided and discussed within the article

Conflicts of Interest

The authors declare that they have no conflicts of interest

References

[1] HHuang andAV Savkin ldquoOptimal path planning for a vehiclecollecting data in a wireless sensor networkrdquo in Proceedings ofthe Chinese Control Conference (CCC rsquo16) pp 8460ndash8463 IEEEJuly 2016

[2] O D Jerew and N A Bassam ldquoEnergy aware and delay-tolerant data gathering in sensor networks with a mobile sinkrdquoin Proceedings of the MEC International Conference on Big Dataand Smart City (ICBDSC rsquo16) pp 1ndash5 IEEE March 2016

[3] O Jerew K Blackmore andW Liang ldquoMobile Base Station andClustering to Maximize Network Lifetime in Wireless SensorNetworksrdquo Journal of Electrical and Computer Engineering vol2012 Article ID 902862 13 pages 2012

[4] Z Xu W Liang and Y Xu ldquoNetwork lifetime maximizationin delay-tolerant sensor networks with a mobile sinkrdquo inProceedings of the 2012 IEEE 8th International Conference onDistributed Computing in Sensor Systems (DCOSS rsquo12) pp 9ndash16IEEE 2012

[5] M Zhao andY Yang ldquoBounded relay hopmobile data gatheringin wireless sensor networksrdquo IEEE Transactions on Computersvol 61 no 2 pp 265ndash277 2012

[6] R C Shah S Roy S Jain and W Brunette ldquoData MULEsModeling a three-tier architecture for sparse sensor networksrdquoin Proceedings of the Sensor Network Protocols and Applications(SNPA rsquo03) pp 30ndash41 IEEE 2003

[7] A Chakrabarti A Sabharwal and B Aazhang ldquoUsing pre-dictable observer mobility for power efficient design of sensornetworksrdquo in Proceedings of the International Conference onInformation Processing in Sensor Networks pp 129ndash145 2003

[8] P Baruah and R Urgaonkar ldquoLearning-enforced time domainrouting to mobile sinks in wireless sensor fieldsrdquo in Proceedingsof the 29th Annual IEEE International Conference on LocalComputer Networks (LCN rsquo04) pp 525ndash532 2004

[9] F N Huang T Y Chen S H Chen H W Wei T S Hsuand W K Shih ldquoShareEnergy Configuring Mobile Relays toExtend Sensor Networks Lifetime Based on Residual Powerrdquoin Proceedings of the International Conference on CollaborationTechnologies and Systems (CTS rsquo16) pp 478ndash484 IEEE Oct2016

[10] Z Chen L Kang X Li J Li and Y Zhang ldquoConstructing load-balanced Degree-constrained Data Gathering Trees inWirelessSensor Networksrdquo in Proceedings of the 2015 IEEE InternationalConference on Communications (ICC rsquo15) pp 6738ndash6742 IEEEJune 2015

[11] S Gao and H Zhang ldquoEnergy efficient path-constrainedsink navigation in delay-guaranteed wireless sensor networksrdquoJournal of Networks vol 5 no 6 pp 658ndash665 2010

[12] AKansalAA SomasundaraDD JeaMB Srivastava andDEstrin ldquoIntelligent fluid infrastructure for embeddednetworksrdquoin Proceedings of the 2nd International Conference on MobileSystems Applications and Services ( MobiSys rsquo04) pp 111ndash124ACM 2004

[13] MMa and Y Yang ldquoSenCar An energy-efficientdata gatheringmechanism for large-scale multihop sensor networksrdquo IEEETransactions on Parallel and Distributed Systems vol 18 no 10pp 1476ndash1488 2007

[14] E M Saad M H Awadalla M A Saleh H Keshk and R RDarwish ldquoA data gathering algorithm for a mobile sink in large-scale sensor networksrdquo in Proceedings of the 4th InternationalConference on Wireless and Mobile Communications (ICWMCrsquo08) pp 207ndash213 IEEE 2008

[15] X Zhang L Zhang G Chen and X Zou ldquoProbabilisticpath selection in wireless sensor networks with controlledmobilityrdquo in Proceedings of the International Conference onWireless Communications Signal Processing pp 1ndash5 IEEE 2009

[16] G Xing T Wang W Jia and M Li ldquoRendezvous designalgorithms for wireless sensor networks with a mobile basestationrdquo in Proceedings of the 9th ACM International Symposiumon Mobile Ad Hoc Networking and Computing (MobiHoc rsquo08)pp 231ndash240 2008

[17] A Kaswan P K Jana and M Azharuddin ldquoA delay efficientpath selection strategy for mobile sink in wireless sensornetworksrdquo in Proceedings of the 2017 International Conferenceon Advances in Computing Communications and Informatics(ICACCI rsquo17) pp 168ndash173 IEEE 2017

[18] A Kaswan K Nitesh and P K Jana ldquoEnergy efficient pathselection for mobile sink and data gathering in wireless sensornetworksrdquo AEU - International Journal of Electronics and Com-munications vol 73 pp 110ndash118 2017

[19] L Chen X Jianxin X Peng and X Kui ldquoAn energy-efficientand relay hop bounded mobile data gathering algorithm inwireless sensor networksrdquo International Journal of DistributedSensor Networks vol 2015 Article ID 680301 9 pages 2015

[20] Y Chen X Lv S Lu and T Ren ldquoA Lifetime OptimizationAlgorithm Limited by Data Transmission Delay and Hopsfor Mobile Sink-Based Wireless Sensor Networksrdquo Journal ofSensors vol 2017 Article ID 7507625 11 pages 2017

[21] M Mishra K Nitesh and P K Jana ldquoA delay-bound efficientpath design algorithm for mobile sink in wireless sensornetworksrdquo in Proceedings of the 3rd International Conference onRecent Advances in Information Technology (RAIT rsquo16) pp 72ndash77 IEEE 2016

[22] C Cheng and C Yu ldquoMobile data gathering with bounded relayin wireless sensor networkrdquo IEEE Internet of Things Journal pp1ndash16 2018

[23] C Zhu S Wu G Han L Shu and HWu ldquoA tree-cluster-baseddata-gathering algorithm for industrial WSNs with a mobilesinkrdquo IEEE Access vol 3 no 4 pp 381ndash396 2015

[24] J-Y Chang and T-H Shen ldquoAn efficient tree-based powersaving scheme for wireless sensor networks with mobile sinkrdquoIEEE Sensors Journal vol 16 no 20 pp 7545ndash7557 2016

[25] O Jerew and K Blackmore ldquoEstimation of hop count in multi-hop wireless sensor networks with arbitrary node densityrdquoInternational Journal of Wireless and Mobile Computing vol 7no 3 pp 207ndash216 2014

[26] T H Cormen C E Leiserson R Rivest and C Stein Introduc-tion to Algorithms The MIT Press 2001

International Journal of

AerospaceEngineeringHindawiwwwhindawicom Volume 2018

RoboticsJournal of

Hindawiwwwhindawicom Volume 2018


Active and Passive Electronic Components

VLSI Design



Shock and Vibration


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of

Advances inOptoElectronics

Hindawiwwwhindawicom

Volume 2018

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Control Scienceand Engineering

Journal of



Journal ofEngineeringVolume 2018

SensorsJournal of



RotatingMachinery


Modelling ampSimulationin EngineeringHindawiwwwhindawicom Volume 2018


Chemical EngineeringInternational Journal of Antennas and

Propagation




Navigation and Observation


Hindawi

wwwhindawicom Volume 2018

Advances in

Multimedia

Submit your manuscripts atwwwhindawicom


or train Thus the sensor nodes can predict when theBS may move around to send their data [7 8]

(3) Controlled mobility when the BS is mounted on arobot or UAV plane then the direction and speedof the BS can be controlled Many algorithms areproposed to find the tour of the BS in order to achieverequirements such as maximised network lifetimeand data-gathering delay using single and multihoprelays [3 5 9]

Most literature studies the mobility of WSNs to bepredictable or controlled to achieve network performancerequirements (some literature focused on randommobility toimprove network lifetime) This differs from mobile ad hocnetworks in which nodes are assumed to move arbitrarilywhich degrades the network performance by link failures

In this paper we deal with the data gathering problemwith bounded delay time in sensor network using a controlledmobile BS by exploring a balance between the relay hopcount required for data relay and tour length of the mobile BSsuch that the energy consumption is minimized Specificallywe assume themobile BS selected a set of Collection Location(CL) nodes that temporarily cache and forward sensing dataof other nodesThe sensor nodes are grouped into CollectionTrees (CTs) rooted at CLs Each node has to forward its datato the CL within a certain number of relay hops The mobileBS traverses along a close tour and stops at each CL in thetour for data gathering

Our major contributions in this paper are as followsWe propose two efficient heuristic algorithms Adjacent Tree-BoundedHopAlgorithm (AT-BHA) and FarthestNode First-Bounded Hop Algorithm (FNF-BHA) The algorithms findthe tour of the mobile BS consisting of CLs which meet thefollowing criteria (i) the energy consumption among thesensor nodes is balanced in order to prolong nodes lifetime(ii) the total traversal time of the mobile BS on the touris minimized and (iii) the hop count between any sensornode and the CL is bounded by a given value The AT-BHAconstructs adjacent CTs by finding closest nodes to the lastconstructed CL The FNF-BHA uses the farthest node fromthe BS in the construction of CL Load balancing algorithm isproposed to balance the number of nodes in the CTs Finallythe performance of the proposed algorithms is evaluatedthrough experimental simulations The experimental resultsdemonstrate that the proposed algorithms are efficient

The remainder of this paper is organized as followsSection 2 reviews the related work on mobile data gatheringSection 3 presents two algorithms to solve data gatheringwith bounded hop count problem Section 4 evaluates theefficiency of the proposed algorithms through extensivesimulations Finally Section 5 concludes the paper

2 Related Work

There is an extensive range of literature studying the useof mobile BS for data gathering [4 5 9ndash15] Kansal et al[12] combine multihop forwarding with a mobile BS Theyperformed an experimental evaluation for a small sensornetwork assuming that a mobile BS moves back and forth

on a straight line (a fixed path) They employed a directeddiffusion approach to gathering sensed data from the sensornodes beyond the transmission range of the mobile BS

Some existing studies assume that sensor nodes can cachesensing data of other nodes in order to forward it to themobile BS when the BS moves around Gao and Zhang [11]and Xing et al [16] explored the delay requirement for datagathering assuming that the sensor nodes have an ability tocache sensing data of other nodes In [11] sensor nodes sendtheir data through multihop relays to the nodes (subsink)located within the direct-transmission range of the mobileBS The subsink nodes cache data and send it to the mobileBS when it comes within transmission range In [16] arendezvous-based data-gathering approach is proposed inwhich a subset of nodes is chosen as rendezvous points Therole of these points is to buffer and aggregate data originatingfrom sensor nodes When the mobile BS arrives within thetransmission range of rendezvous points the data will beforwarded to the mobile BS Kaswanet et al [17] proposedpath plan algorithm for data collection by selecting set ofrendezvous points The algorithm minimize the tour lengthby reducing the number of rendezvous points by selecting theposition closest to the location of the BS and allows the BS tocollect data frommaximum number of sensors via single hopcommunication This algorithm does not consider a boundedhop count for data collection

Grouping sensor nodes in the sensing field into clustersand collecting data from cluster heads are proposed in [13ndash15]Ma and Yang [13] proposed a heuristic for finding routingpaths for the mobile BS which assumes that the moving pathof the mobile BS consists of a series of line segments Sensornodes closest to each line segment are selected as clusterheads A specified configuration is applied where the mobileBS starts data gathering from the left side of the path movestowards the right side and then comes back to the left sideagain This scheme maximizes network lifetime However itmay cause packet losses at each cluster head if the path lengthto the cluster head is too long since the time required for thecluster head to send the data of cluster nodes to the BS is notconsidered in cluster forming Saad et al [14] proposed pathplanning algorithm for mobile BS that visit a set of clusterheads Nodes are randomly grouped to form clusters and thenode with the largest residual energy is selected as the clusterhead The clusters are merged ensuring that the number ofhops from the cluster head to any cluster node is no morethan two hops The cluster head gathers and buffers clustersensing data The BS visits the cluster heads to gather datathrough single-hop communicationHowever thiswork doesnot consider data gathering delay for the mobile BS Zhang etal [15] extended the work in [16] and regarded cluster headsas the rendezvous point They considered event based datagathering and proposed reduction in data gathering delayby only allowing the mobile BS to visit cluster heads thatgenerate new data

Bounded hop count for data gathering is studied in [45 18 19] Zhao and Yang [5] proposed SPT-DGA and BRH-MDG data gathering algorithms that select set of CL for themobile BS to visit for data gathering Each CL aggregatesthe local data from its affiliated sensors within a certain


























































0

500

1000

1500

2000

2500

3000

3500

4000

4500

Tour

Len

gth

m

3 4 5 62Hop count h




20

40

60

80

100

120

140

160

180M

ax N

umbe

r of N

odes

3 4 5 62Hop count h








3 4 5 62Hop count h

0

20

40

60

80

100

120

140

No

of C

olle

ctio

n Lo

catio

ns




3 4 5 62Hop count h

94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J







3 4 5 62Hop count h

0

20

40

60

80

100

120

140

Max

Dist

ance

m


0 50 100 150 200


0

50

100

150

200






0 50 100 150 200


0

50

100

150

200


0 50 100 150 200


0

50

100

150

200




0

500

1000

1500

2000

2500

Tour

Len

gth

m






0

10

20

30

40

50

60

70

No

of C

olle

ctio

n Lo

catio

ns








0

20

40

60

80

100

120

Max

Num

ber o

f Nod

es







94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J




0

10

20

30

40

50

60

70

80

90

Max

Dist

ance

m







0

50

100

150

200

250

300

350

400

450

Num

ber o

f Occ

urre

nce



5 Conclusion




Data Availability




References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



























































0

500

1000

1500

2000

2500

3000

3500

4000

4500

Tour

Len

gth

m

3 4 5 62Hop count h




20

40

60

80

100

120

140

160

180M

ax N

umbe

r of N

odes

3 4 5 62Hop count h








3 4 5 62Hop count h

0

20

40

60

80

100

120

140

No

of C

olle

ctio

n Lo

catio

ns




3 4 5 62Hop count h

94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J







3 4 5 62Hop count h

0

20

40

60

80

100

120

140

Max

Dist

ance

m


0 50 100 150 200


0

50

100

150

200






0 50 100 150 200


0

50

100

150

200


0 50 100 150 200


0

50

100

150

200




0

500

1000

1500

2000

2500

Tour

Len

gth

m






0

10

20

30

40

50

60

70

No

of C

olle

ctio

n Lo

catio

ns








0

20

40

60

80

100

120

Max

Num

ber o

f Nod

es







94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J




0

10

20

30

40

50

60

70

80

90

Max

Dist

ance

m







0

50

100

150

200

250

300

350

400

450

Num

ber o

f Occ

urre

nce



5 Conclusion




Data Availability




References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia





3 4 5 62Hop count h

0

20

40

60

80

100

120

140

No

of C

olle

ctio

n Lo

catio

ns




3 4 5 62Hop count h

94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J







3 4 5 62Hop count h

0

20

40

60

80

100

120

140

Max

Dist

ance

m


0 50 100 150 200


0

50

100

150

200






0 50 100 150 200


0

50

100

150

200


0 50 100 150 200


0

50

100

150

200




0

500

1000

1500

2000

2500

Tour

Len

gth

m






0

10

20

30

40

50

60

70

No

of C

olle

ctio

n Lo

catio

ns








0

20

40

60

80

100

120

Max

Num

ber o

f Nod

es







94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J




0

10

20

30

40

50

60

70

80

90

Max

Dist

ance

m







0

50

100

150

200

250

300

350

400

450

Num

ber o

f Occ

urre

nce



5 Conclusion




Data Availability




References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia



0 50 100 150 200


0

50

100

150

200


0 50 100 150 200


0

50

100

150

200




0

500

1000

1500

2000

2500

Tour

Len

gth

m






0

10

20

30

40

50

60

70

No

of C

olle

ctio

n Lo

catio

ns








0

20

40

60

80

100

120

Max

Num

ber o

f Nod

es







94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J




0

10

20

30

40

50

60

70

80

90

Max

Dist

ance

m







0

50

100

150

200

250

300

350

400

450

Num

ber o

f Occ

urre

nce



5 Conclusion




Data Availability




References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




0

20

40

60

80

100

120

Max

Num

ber o

f Nod

es







94

95

96

97

98

99

100

Max

Ene

rgy

Con

sum

ptio

n of

CL

Per S

enso

r J




0

10

20

30

40

50

60

70

80

90

Max

Dist

ance

m







0

50

100

150

200

250

300

350

400

450

Num

ber o

f Occ

urre

nce



5 Conclusion




Data Availability




References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




Data Availability




References





























RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia




RoboticsJournal of




VLSI Design



Shock and Vibration







Journal of



Volume 2018



Volume 2018


Journal of




SensorsJournal of



RotatingMachinery





Propagation






Hindawi


Advances in

Multimedia


Delay Tolerance and Energy Saving in Wireless Sensor...

Documents

Transcript of Delay Tolerance and Energy Saving in Wireless Sensor...