A Comparative Study of Deployment Strategies in Wireless Sensor Networks A Comparative

International Journal of Engineering, Management & Sciences (IJEMS) ISSN-2348 –3733, Volume-2, Issue-5, May 2015

A Comparative Study of Deployment Strategies in Wireless Sensor Networks Krishna P. Sharma, T.P. Sharma  Abstract— Wireless Sensor Networks (WSNs) generally are deployed in hostile and remote areas. The major challenge in such ill-disposed areas is to find a tradeoff between desired and contrary requirements for the lifetime, coverage and cost with limited available resources. In this paper, we present the study of various sensor network deployment techniques and their classifications. We mainly classified deployment techniques on the basis of Region of Interest (ROI) characteristics and mathematical approach used by them. On the basis of ROI characteristics, there are three types of deployment approaches like deterministic, random and self-deployment. We further categorize deployment techniques on the basis of four mathematical approaches used to solve the deployment problem like Genetic Algorithms (GAs), Computational Geometry, Artificial Potential Field (APF), and Particle Swarm Optimization (PSO). In the last an extensive comparison of deployment techniques on the basis of discussed metrics is presented with their strength and limitations.

Index Terms—Wireless Sensor Networks (WSNs); Genetic Algorithms (GAs); Artificial Potential Field (APF); Virtual forces, Computational Geometry; Particle Swarm Optimization (PSO)

I. INTRODUCTION The rapid advancement in communication, Micro-Electronic Mechanical Systems (MEMS), robotics and chip technologies facilitated the small and inexpensive wireless sensor nodes. Sensor nodes are resource constraint battery powered devices, embedded with limited communication, processing, sensing and storage capabilities.

The modern sensors are wireless, with or without locomotive capabilities and can sense humidity, temperature, pressure, vibrations and other physical parameters. Due to sensing versatility and low cost, wireless sensor nodes are vastly used in almost every field for various applications [1, 2] like environmental monitoring, industrial real-time monitoring and automation, target tracking in military operations, traffic surveillance and control, continuous health monitoring, etc.

WSNs are application specific, i.e. WSNs behaviors and performance requirements vary according to applications.

WSNs support variety of applications in the field of military surveillances, home automation, biological monitoring, environmental monitoring, habitat monitoring, etc. Different applications have its own challenges in terms of coverage, connectivity and quality of data needed which Manuscript received February 20, 2015 Krishna P. Sharma, Computer Science and Engineering, National Institute of Technology, Hamirpur, Himachal Pradesh, India, T.P. Sharma, Computer Science and Engineering, National Institute of Technology, Hamirpur, Himachal Pradesh, India

1

are heavily affected by the deployment strategy used to construct the network. There are many proposed deployment approaches for optimizing coverage, connectivity and lifetime of WSNs. But the optimal node placement in WSNs is a very challenging task which has been proven to be NP-Hard for most of the formulations of sensor deployment [3-4]. To tackle such complexity, several heuristics have been proposed to find sub-optimal solutions [3]. In case of some applications like home automation, industrial automation, etc. sensing field is approachable, where the best utilization point within the ROI can be calculated for deterministically placing the sensors. This deterministic placement of sensor nodes constructs a high performance network. A lot of work exists in [22-23, 25, 28-29] which consider mobile sensors and propose various approaches to relocate the randomly deployed sensors to achieve high coverage with minimum energy consumption. In such approaches, mobile nodes have to move to best utilizing positions in order to enhance effective coverage with minimum movement and using less number of sensors. The use of mobile nodes in WSNs adds more dimensions and challenges in terms of the energy consumed in movement for relocation and executing the relocation algorithm in a distributed environment. Schemes presented in [7, 25, 28-29] consider these challenges and propose various relocation algorithms for maximizing the coverage.

This paper categorizes various strategies of sensor placement with varying characteristics. It explains some performance metrics for evaluating the performance of deployed networks. Further, paper contrasts a number of deployment approaches for various kinds of applications and highlights their strengths and limitations.

The paper is organized as follows. The next section gives the description of performance metrics and some models to evaluate deployed networks. In section III, we present the different classifications of deployment techniques. Section IV, gives an extensive comparison of classified deployment techniques on the basis of deployment performance metrics and models followed by a conclusion.

II. PERFORMANCE COMPARISON METRICS There are many approaches for deploying WSNs and the appropriate selection can be made according to the application. The performance of a sensor network can be measured on the basis of various parameters like coverage, connectivity, lifetime of the network, time to converge, etc.

This section gives some evaluation metrics to evaluate the performance of deployed networks.

www.alliedjournals.com

A Comparative Study of Deployment Strategies in Wireless Sensor Networks A. Coverage Coverage area [5] is an area A which is covered by N sensors such that for every point x in A the di(x)<Rs (Sensing Range), where the di(x) is the Euclidian distance of sensor i from x for i,(1≤i≤ N). When every point x in A is covered by at least k sensors or there are at least k sensors at a distance less than Rs from x then this is called k-coverage.

B. Connectivity The primary task of a sensor network is to detect or sense some physical quantity within the sensing range and transfer the sensed data to the sink through any connected route. For successful transfer of sensed data, every node of the network should be connected to the sink by any possible route of one or multiple hops. A sensor network is connected if there exist at least one path between each pair of nodes [5]. If in a network at least k(k ≥1) path exists between any pair of nodes, then the network is called k-connected and k-connectivity in the network.

C. Uniformity Sensors closer to the sink generally participate more than others during data transmission. Therefore, to keep nearly the same energy level throughout the entire network, more numbers of nodes can be placed near to the sink. The uniform node density decreases the chances of congregating and coverage holes. The uniformity can be defined as the average local standard deviation of the distance between nodes [7]: U 

and

1 N

N

U

1

Ki

 (d j 1

i, j

2  M i )  

III. CLASSIFICATION AND EVALUATION The deployment phase of the sensor network influences the overall performance and life span of a network.

According to the literature studied, the deployment techniques can be classified in various ways. In this section, the deployment techniques are classified on the basis of field characteristics and mathematical approach used by them. On the basis of field characteristics, the deployment techniques are classified as deterministic, random and self deployment.

Whereas, on the basis of mathematical background, the deployment techniques are categorized in four ways like Genetic Algorithm (GA) based, Computational Geometry (CG) based, Artificial Potential Field (APF) based and Particle Swarm Optimization (PSO) based techniques. The classification is shown in Fig.1. However, the maximum deterministic, random and self–deployment techniques use the above defined mathematical approaches.

A. Classification based on sensing field characteristics WSNs currently are deployed on land, underground, underwater, within machine, etc. for a huge range of its applications. The different environments have their own challenges and constraints [38].

i

i 1

 1 Ui    Ki 

F. Obstacle Adaptability WSNs have a huge list of targeted applications with varying characteristics of ROI. For most of the applications, field cannot be considered always 2D or 3D plane surface as there may be many obstacles in the ROI. During the deployment, it is necessary to consider obstacles in the field and modeling their effect on the performance of deployed networks. Many authors consider the obstacles during deployment and propose various approaches to handle the obstacle in the field.

(1)

where, N is the total number of nodes, Ki is the number of neighbors of the th node, di,j is the distance between th and th nodes, and Mi is the mean of inter-modal distances between the ith node and its neighbors.

D. Time In case of some time-critical applications such as search-and-rescue and disaster recovery operations, time spent in deployment of a network is very important [6]. The convergence time is an important parameter to improve in case of mobile WSNs. The time for deploying a working mobile sensor network depends upon complexity of relocation algorithms. The total time elapsed is defined here as the time elapsed until all the nodes reach their final positions.

E. Energy Energy utilization is an important issue in WSNs. Energy is a critical factor to consider during the development of relocation algorithms for MWSNs. Mainly the energy is consumed in computation, communication and movement of mobile sensors during the relocation process. The energy consumption during relocation is subject to improve because it is very high as compare to energy consumed in general processing of WSNs.

2

Deterministic Deployment Deterministic deployment of sensors is possible where region of interest is accessible. In deterministic deployment, the calculation of best utilizing locations for enhancing the network performance is possible. This calculation is based on various optimization theories like linear and non-linear programming, GAs, PSO, etc. In case of deterministic deployment, some time placement of sensors follows certain predefined patterns like hexagonal grid, triangular grid, rectangular grid, etc. Many approaches have been proposed, which consider the accessible ROI where the deterministic deployment is possible. In [9], a deployment approach is proposed which uses the maximum multi-overlapping domains of target points and GAs to reduce the deployment cost and improve the network coverage. In [10], the author proposes a method for deployment based on grid scan.

Scheme proposed in [12-15] are also the deterministic deployment techniques. The technique in 21] uses the GAs with virtual force. In [14], the author proposed a genetic based algorithm called MOGA (Multi Objective Genetic algorithm) for deploying N sensors in a 2D field for achieving a high degree of coverage. Where, the approach in [15] also based on Genetic Algorithm and is an extension of [14]. In this approach, the author considers three specific surveillance scenarios where each scenario had its own set of objectives according to the surveillance required. In [14-15],

www.alliedjournals.com

International Journal of Engineering, Management & Sciences (IJEMS) ISSN-2348 â&#x20AC;&#x201C;3733, Volume-2, Issue-5, May 2015 approaches are for area coverage and use the binary sensing considers the concept of VD for improving the coverage. model. The approach in [13] is for point coverage with the Given approaches use the binary sensing model. The objective of covering a finite set of target points with some technique in [22], uses both static and mobile sensors. The predetermined number of sensors. The approach in [21], author proposed a distributed bidding protocol to assist the consider 3D plane to perform the optimization work for movement of mobile sensors. The approach in [23], first WSNs. In this approach, the author describes a novel discovers the coverage hole and then proposed protocol force-driven evolutionary approach for solving calculate the target position for sensor to move. In [24], the multi-objective 3D deployment problems in differentiated author proposed two approaches based on the centroid of VD. WSNs. In [7], the author considers the same deployment problem as in [22-23] and [24] but with the objective of minimizing the Random Deployment energy consumption. The techniques in [28, 30-31] are based In many applications like battlefield surveillance, habitat on virtual force where the sensor nodes, ROI and obstacles in monitoring, disaster management, etc. the sensing field is ROI are the sources of repulsive and attractive force. In case hostile. For such applications, sensor nodes are thrown of self-deployment, maximum techniques uses the concept of randomly because deterministic deployment is not possible. virtual force, but some approaches like in [34-35] uses the Some researchers improved and utilized the network energy concept of PSO to improve the coverage and connectivity by scheduling those nodes which overlap their coverage. In with minimum use of energy consumption.

[17] and [18], the authors proposed approaches to schedule nodes of high density regions in order to enhance network lifetime and maintain effective coverage and connectivity.

There are many random deployment strategies to meet the requirement. The techniques proposed in [17-20] are the Deterministic random deployment techniques which are mathematically based on GAs. The author mainly concerns on the scheduling of randomly deployed sensors in [17]. In this scheme, the Deployment Random author introduced a Non-dominated Sorting Genetic Algorithm (NSGA-II) to enhance the network lifetime and Self-Deployment maintain the proper coverage by keeping the switch on minimum number of sensors. In [18], the similar approach as [17] is proposed, but this approach mainly focused on the (a) Deployment classification on the basis of field improvement of selecting dynamic sensor nodes which were Characteristics first introduced by Burne in [19]. The approach reduces the search space of GAs by using a Hopfield network (NH).

Schemes in [19] and [20] are also used the principle of GAs Genetic Algorithm with VD to improve the network performance. In [20], the Based sensing range adjustable sensors are considered, where the coverage is maximized by adjusting the sensing range in Computational order to cover their respective Voronoi cell. The approach in Geometry Based [8] is based on Computational Geometry (CG) uses Delaunay Deployment Triangle particularly. This scheme uses static sensors in 2D Potential Field Based plane with some obstacle in ROI. The Computational Geometry is used to identify the large coverage hole and then Particle Swarm it covers by the placement of sensors.

Optimization Based Self-Deployment Due to the advancement of MIMS and robotics (b) Deployment classification on the basis of technology, now the mobile sensors are possible. Mobile mathematical background sensors have the capability to move in some limited region after throwing them randomly from any source. Sensors are battery powered, resource constrained devices which are generally being deployed in an environment where the replacement of their battery and sensor nodes are not Fig. 1. Classifications of various deployment techniques possible. Random deployment is a non-uniform deployment where the uniformity is achieved by making some or all B. Classification Based on Mathematical Background nodes mobile. Therefore, some algorithms have developed to Many authors solve this optimization problem with the rearrange the randomly deployed sensor nodes. The main help of mainly four mathematical techniques: Genetic goal of relocation algorithm is to minimize the energy Algorithms (GAs), Computational Geometry (CG), Artificial consumption in traveling distances and communication to Potential Field (APF), and Particle Swarm Optimization achieve the best suited position for sensors. Some techniques (PSO). in [22-24] and [7] are based on Computation Geometry which

3

www.alliedjournals.com

A Comparative Study of Deployment Strategies in Wireless Sensor Networks Genetic Algorithms (GAs) Based Genetic algorithms (GAs) are a type of evolutionary, optimization algorithms based on the mechanics of natural selection and genetics. Such type of optimization algorithms becomes popular after they introduced by Barricelli in 1957 [11]. GAs have been used for solving optimization problems in various fields such as computer networking, industrial engineering and machine learning. GAs are heuristic search algorithms that come from the idea of natural evolution and use the concept of inheritance, mutation, chromosomes and crossover. Gas are effective in combinatorial and multi-objective optimization problems, in which deterministic optimization methods are not applicable.

In [14], the author proposes a Multi Objective Genetic Algorithm (MOGA) for deploying N static sensor nodes for achieving maximum coverage and lifetime by using the minimum number of sensors. The author assumes 2D type of region of interest (ROI), homogeneous sensors with the binary sensing model. In [15], the authors extends their work by using the same MOGA of [14], but applied to three specific surveillance scenarios. Each scenario had its own set of objectives according to the surveillance required. In the first scenario, considered objectives are maximizing coverage, minimizing the number of sensors and maximizing the distance between deployed sensors. The second scenario is for barrier coverage while the third is for area coverage.

Both the last scenarios require maximizing coverage while minimizing the number of sensors for deploy. The approaches in [14] and [15] are flexible, which can be applied to the scenarios with different set of objectives. The scheme proposed in [13] is for point coverage, where the objective is to cover a finite set of target points with the predetermined number of sensors. The algorithm is evaluated with six target points and five deployment points which give the optimal solution by generating all possible combinations. The algorithm in [12] is the extension of the approach defined in [13] which uses the probability sensing model. It uses the same GAs for finding the optimal solution as used in [13] but added new constrained for probability sensing in the formulated optimal problem. The approach in [16] is similar to [14] and [15] with the main objective of maximizing the coverage area in 2D plane. The approach uses the binary sensing model and the heterogeneous sensors with three different sensing ranges. The approach uses the GAs for solving the Maximum Coverage Sensor Deployment Problem (MCSDP). Candidate solution in the phenotype space is converted to the genotype space using an integer coded chromosome as in [14] and [15] with the difference that the coordinates of the sensors which belongs to same sensor type in a chromosome are contiguous.

Computational Geometry (CG) Based The computational geometry (CG) is basically used to solve the coverage problem in the sensor network deployment. The sensor nodes are assumed as a point in 2D and 3D Region of Interest. The region is divided generally in two geometric structures one is Voronoi Diagram (VD) and another is Delaunay Triangulation (DT). The VD is a fundamental construct defined by a discrete set of points [36].

In 2D plane, the discrete points (site) partition the plane into various convex polygons called Voronoi cell for each point

4

(site) such that all points inside a polygon are closest to only one site. Another structure is a Delaunay triangulation [8] which is directly related to Voronoi diagrams. The Delaunay Triangulation is constructed by connecting the sites in the Voronoi diagram whose polygons share a common edge.

In [22], the approach uses mobile as well as static sensor, where the mobile sensor move from a dense area to sparse area for better coverage. The author proposed a distributed bidding protocol to assist the movement of mobile sensors.

The proposed scheme in [23] is a distributed self-deployment protocols for mobile sensors. The approach uses Voronoi diagrams to discover the coverage holes and design three movement-assisted sensor deployment protocols, VEC (VECtorbased), VOR (VORonoi-based), and Minimax based on the principle of moving sensors from densely deployed areas to sparsely deployed areas. The approach depends on locally constructed Voronoi cells as in [22]. The target location in each iteration of sensor is a point inside the local Voronoi cell called the Minimax point. In [24], two distributed and iterative deployment algorithms are proposed, namely the Centroid and the Dual-Centroid with the objective of maximizing the area coverage of initially deployed MWSNs. The algorithm improves the coverage in each iteration, which is computed by locally constructed Voronoi cell of sensors in the same way as in [23] and [22].

The results of both algorithms are compared with the approach in [23]. The result shows that both algorithms have better performance than of Minimax presented in [23].

Artificial Potential Field (APF) Based Artificial Potential Field (APF) based techniques were first introduced in the field of Robotics [37]. The idea in the scheme is similar to the equilibrium of molecules, which minimizes molecular electronic energy and inter-nuclear repulsion. Each molecule determines its own lowest energy point in a distributed manner and its resulting spacing from the other particles is almost the same. There are various techniques which use the concept of virtual force.

In [26], a virtual force based approach is used to optimize the coverage area by the mobile sensors. The algorithm considers ROI with obstacle where the sensor node and obstacle create some virtual forces. In [27], the author presents sensor network architecture, called active sensor network (ASN), where in the traditional sensor network some network friendly mobile robots are inserted. These robots carry the sink nodes, which relocates by the use of the potential field algorithm. The author in [28] proposes a virtual force based deployment and target localization scheme to enhance the coverage of randomly deployed mobile sensors of similar type. In the scheme proposed in [29], the author presents an algorithm based on Target Involved Virtual Force Algorithm (TIVFA) for self-deployment in the context of target tracking. The presented algorithm dynamically adjusts sensor network configuration according to the terrains, intelligence and those detected maneuvering targets for improving coverage and detection probability. The approach in [30] is the improvement of the virtual force algorithms by applying a back-off delay time. In [31], the author proposes the similar king of algorithm based on repulsive and attractive forces.

But in [31], the author considers mobile and static sensors.

www.alliedjournals.com

International Journal of Engineering, Management & Sciences (IJEMS) ISSN-2348 –3733, Volume-2, Issue-5, May 2015 Some sensors carrying robots are inserted in the ROI, aiming by them, namely, Genetic Algorithm based, Computational to achieve certain goal to detect an activity happening in the Geometry based, Artificial Potential Field Based, and ROI by at least four of the deployed robotic sensors. The Particle Swarm Optimization based. We presented a brief robotic sensors then transmit the sensed data to a fixed sink discussion of some parameters and models to evaluate the node. In [26], only repulsive force is used where in [31] both performance of deployed networks. The comparative repulsive and attractive forces are acting on sensors. discussion of deployment techniques belongs to each category was presented.

Particle Swarm Optimization (PSO) Based Particle Swarm Optimizations (PSOs) are based on Swarm REFERENCES Intelligence, which is a branch of Artificial Intelligence (AI) [1] V. Jelicic, T. Razov, D. Oletic, M. Kuri, and V. Bilas, “Maslinet: A and first introduced by R. Eberhart et. al. in 1995 [39]. In this wireless sensor network based environmental monitoring system,” in Proc. 34th International Convention on Information and approach, there are some interacting agents organized in Communication Technology, Electronics and Microelectronics, 2011, small societies, called swarms, which exhibit traits of pp. 150– 155. intelligence, such as the ability to react to environmental [2] Stankovic, John A., Wood, Anthony D., and He, Tian. “Realistic threats and decision making capacities. The individual in the Applications for Wireless Sensor Networks,” Theoretical Aspects of Distributed Computing in Sensor Networks, Springer Verlag, 2010. group finds the best solution or particles and keep it into [3] A. Efrat, S. Har-Peled, J.S.B. Mitchell, “Approximation algorithms for memory as an experience. That experience, then two optimal location problems in sensor networks,” in: Proceedings of communicates to the part of the swarm to directing the the 3rd Broadnets, Boston, Massachusetts, Oct 2005. movement towards the search space region where it is more [4] S. Poduri, S. Pattem, B. Krishnamachari, G.S. Sukhatme, “Sensor network configuration and the curse of dimensionality,” in: likely finding the optimal solution. There are various Proceedings of the 3rd IEEE Workshop on Embedded Networked techniques [25, 32-33] in WSNs, which uses PSOs for Sensors, Cambridge, MA, May 2006. network deployment. The PSOs also used in base station [5] Sonali Sen Baidya, C.K.Bhattacharyya “Coverage and Connectivity in Wireless Sensor Networks: Their trade-offs,” Sixth International positioning, node localization, data aggregation, and energy conference on Sensing Technology (ICST), IEEE 2002. aware clustering.

[6] A. Howard, M. J. Mataric´, and G. S. Sukhatme (2002). “An incremental self deployment algorithm for mobile sensor networks,” In [32], author’s main objective is to deploy a finite Robots, 13( 2), 113–126. number of N -sensors of similar type in 2D ROI to reduce [7] Autonomous Nojeong Heo and Pramod K. Varshney (2005). “Energy-Efficient cost by clustering method based on distance based energy Deployment of Intelligent Mobile Sensor Networks” IEEE Transaction on System, Man, and cybernetics-Part A: System and model. Initially, sensors are deployed randomly in sensing 35( 1). field, and then the PSO-based algorithm executes on a [8] Humans, C.-H. Wu, K.-C. Lee, and Y.-C. Chung, “A delaunay triangulation powerful base station centrally and transmits optimized based method for wireless sensor network deployment,” in Proc. 12th ICPADS’06, 2006. location information to each randomly deployed sensor node.

[9] X. He, X. L. Gui, and J. An (2010). “A Deterministic Deployment The study in [24], considers the main objective is to Approach of Nodes in Wireless Sensor Networks for Target maximize the area coverage of WSNs composed by a finite Coverage,” Journal of XiAn Jiaotong University, 44(6)， 6-10. number of homogeneous static sensors in a 2D ROI. The [10] X. M. Guo, C. J. Zhao, X. T. Yang et al.(2012). “A Deterministic Sensor Node Deployment Method with Target Coverage Based on algorithm uses the same particle encoding in PSO-Grid as in Grid Scan”. Chinese Journal of Sensors and Actuators, 25(1), [32] and used a Voronoi diagram to calculate the coverage 104-109 area. In [33], the author uses a parallel PSO algorithm for [11] Barricell and N. Aall (1957). "Symbiogenetic evolution processes realized by artificial methods". Methods, 143-182. optimizing the energy used by sensors to track targets and [12] B. Carter and R. Ragade(2009). “A probabilistic model for the maximizing the coverage. The author optimizes the deployment of sensors”, IEEE Sensors Applications Symposium, 7–12. consumed energy in target detection by turning off those [13] B. Carter and R. Ragade (2008). “An extensible model for the deployment of non-isotropic sensors” in Proc. IEEE Sensors sensors which are far from the target. The schemes proposed Applications Symposium (SAS’08), 22–25. in [34] are also based on PSO but for mobile sensors. The [14] D. B. Jourdan and O. L. de Weck (2004). “Layout optimization for a scheme proposed in [34] used PSO to relocate the mobile wireless sensor network using a multi-objective genetic algorithm,” in Proc. IEEE Semi-annual Vehicular Technology Conference-Spring sensors in ROI in order to optimize the coverage area. The 5, 2466–2470. algorithm runs after the random deployment of the sensor [15] (VTC’04-Spring), D. B. Jourdan and O. L. de Weck,(2004). “Multi-objective genetic node to maximize the coverage with minimum movement. In algorithm for the automated planning of a wireless sensor network to monitor a critical facility,” in Proc. SPIE Defense and Security [35], the author proposed and algorithm called VFCPSO Symposium, 565–575. which uses the concept of Virtual Force (VF) and [16] Y. Yoon and Y. H. Kim (2013). “An efficient genetic algorithm for co-evolutionary particle swarm optimizations (CPSO). The maximum coverage deployment in wireless sensor networks,” IEEE Transactions on Cybernetics. algorithm updates the evolving velocity of each candidate solution and different, cooperating swarms by the use of [17] J. Jia, J. Chen, G. Chang, J. Li and a. Y. Jia (2007). "Coverage Optimization based on Improved NSGA-II in Wireless Sensor virtual forces.

Network," IEEE International Conference on Integration Technology, IV. CONCLUSION In the paper, we classified various deployment techniques for Wireless Sensor Networks and Mobile Wireless sensor Networks in two ways. First, we classified the deployment techniques on the basis of ROI characteristics, namely, deterministic, random, and self. We proposed second classification on the basis of mathematical background used

5

614-618, 2007.

[18] X. Wang, S. Wang and D. Bi (2007). "Dynamic sensor nodes selection strategy for wireless sensor networks," International Symposium on Communications and Information Technologies, 1137-1142.

[19] R. A. Burne, A. L. Buczak and a. Y.C.Jin (1999). "A self-organizing, cooperative sensor network for remote surveillance:current result," 13th Annua lInternational Symposium on AeroSense Conference–Unattended Ground Sensor Technologies and Applications, 238-248.

www.alliedjournals.com

A Comparative Study of Deployment Strategies in Wireless Sensor Networks [20] H.-C. Jang and a. H.-C. Len (2007). "A Voronoi dEtection Range Adjustment (VERA) approach for energy saving of wireless sensor networks," International Conference on Parallel and Distributed Systems, 2, 1-7.

[21] L.-C. Wei, C.-W. Kang and J.-h. Chen (2009). "A force-driven evolutionary approach for multi-objective 3D differentiated sensor network deployment," IEEE 6th International Conference on Mobile Adhoc and Sensor Systems, 983- 988.

[22] Guiling (Grace) Wang, Guohong Cao, Piotr Berman, and Thomas F. La Porta (2007. Fellow, IEEE, ” Bidding Protocols for Deploying Mobile Sensors”, IEEE Transaction on mobile computing, 6.

[23] G. Wang, G. Cao, and T. La Porta (2004). “Movement-assisted sensor deployment,” in Proc. Annual IEEE International Coneference on Computer Communications (INFOCOM’04), 2469–2479.

[24] H.-J. Lee, Y.-h. Kim, Y.-H. Han, and C. Y. Park (2009).

“Centroid-based movement assisted sensor deployment schemes in wireless sensor networks,” in Proc. IEEE 70th Vehicular Technology Conference-Fall , 1–5.

[25] N. Azlina, Bt. Ab Aziz, A. W. Mohemmed and M.Y. Alias(2009). “A Wireless Sensor Network Coverage Optimization Algorithm Based on Particle Swarm Optimization and Voronoi Diagram” Proceedings of the 2009 IEEE International Conference on Networking, Sensing and Control, Okayama, Japan.

[26] A. Howard, M. Mataric and G. Sukhatme (2002). "Mobile Sensor Network Deployment using Potential Fields: A Distributed, Scalable Solution to the Area Coverage Problem," in the 6th International Symposium on Distributed Autmomous Robotics System, Japan.

[27] W. Sheng, G. Tewolde and S. Ci (2006). "Micro Mobile Robots in Active Sensor Networks: Closing the Loop," in Proceeding of IEEE International Conference on intelligent Robots and Systems (RSJ), Beijing, China.

[28] Yi. Zou and K. Chakrabarty (2003). "Sensor Deployment and Target Localization Based on Virtual Fources," in IEEE Societies Twenty-Second Annual Joint Conference of the IEEE Computer and Communications (INFOCOM), San Fransisco, USA. [29] S. Li, C. Xu, W. Pan and Y. Pan (2005). "Sensor deployment optimization for detecting maneuvering targets," in 8th International Conference on Information Fusion, PA, USA. [30] T. Wong, T. Tsuchiya and T. Kikuno (2004). "A self-organizing technique for sensor placement in wireless micro-sensor networks," in 18th International Conference on Advanced Information Networking and Applications, Fukuoka, Japan.

[31] D. O. Popa, H. E. Stephanou, C. Helm, and A. C. Sanderson(2004).

“Robotic deployment of sensor networks using potential fields,” in Proc. IEEE International Conference on Robotics and Automation (ICRA ’04), 1, 642–647.

[32] W. Xiaoling, S. Lei, Y. Jie, X. Hui, J. Cho, and S. Lee (2005). “Swarm based sensor deployment optimization in ad hoc sensor networks,” in Proc. International Conference on Embedded Software and Systems (ICESS’05). Springer,. 533–541.

[33] X. Wang, J. Ma, S. Wang and D. Bi (2009). "Distributed Energy Optimization for Target Tracking in Wireless Sensor Networks," IEEE Transactions on Mobile Computing, 9(1).

[34] X. Bai, S. Li, C. Jiang and Z. Gao(2009). "Coverage Optimization in Wireless Mobile Sensor Networks," 5th International Conference on Wireless Communications, Networking and Mobile Computing, 1-4.

[35] X. Wang, S. Wang and a. J.-J. Ma (2007). "An Improved Co-evolutionary Particle Swarm Optimization for Wireless Sensor Networks with Dynamic Deployment," Sensors, 7, 354–370.

[36] F. Aurenhammer (1991). “Voronoi Diagrams -A Survey of a Fundamental Geometric Data Structure” ACM Computing Surveys 23, 345-405.

[37] O. Khatib (1986). “Real-time obstacle avoidance for manipulators and mobile robots,” International J. Robotics Research, 5(1), 90–98.

[38] D. Puccinelli and M. Haenggi (2005). “WSN: Applications& Challenges,” CAS Magazine.

[39] R. Eberhart and J. Kennedy (1995). “A new optimizer using particle swarm theory,” in Proc. 6th International Symposium on Micro Machine and Human Science, 39–43.

Teek Parval Sharma is an Associate Professor at National Institute of Technology, Hamirpur (India). He has done his Ph.D from Indian Institute of Technology, Roorkee (Electronics and Computer Engineering Department), India in the area of Wireless Sensor Networks. He has published numerous high quality research papers in International\ National journals and conferences, and has also contributed in various books of standard international publishers. His research interest includes distributed systems, wireless sensor networks, mobile ad hoc networks and wireless networks.

Krishna Pal Sharma received his B.Tech degree in Computer Science and Engineering from Uttar Pradesh Technical University, Lucknow in 2005 and M.Tech in Computer Science and Engineering from Graphic Era University, Dehradun (UK), India in 2010 respectively. Currently, he is working towards his PhD in Sensor Adhoc Network from the Department of Computer Science and Engineering, National Institute of Technology Hamirpur (H.P), India. His research interest includes computer networks and Wireless Sensor Adhoc Networks.

6

www.alliedjournals.com