Chapter 91

Mobile Sink with Mobile Agents:

Effective Mobility Scheme for Wireless Sensor Network

• Rachana Borawake-Satao
  Smt. Kashibai Navale College of Engineering (Savitribai Phule Pune University), Pune, India
• Rajesh Shardanand Prasad
  Computer Engineering Department, NBN Sinhgad School of Engineering (Savitribai Phule Pune University), Pune, India

ABSTRACT

Wireless Sensor Networks (WSN) has been drawing considerable attention and discussion in recent years due to its potential applications in various fields. In modern applications for future internet the MSN (Mobile Sensor Network) is a key factor. Mobility allows the applications of Wireless Sensor Network to be compatible with IoT (Internet of Things) applications. As mobility enhances capability of the network it also affects the performance of the network at each layer. In recent years the various methodologies are proposed to handle mobility. Most of them use mobility for efficient data collection in WSNs. The purpose of this paper is to study effects of mobility on various performance parameters of the network and to explore the effective techniques to handle mobility in network. This paper proposes Mobile Sink with Mobile Agent mobility model for WSN which will increase the lifetime of the network using sink and agent node mobility.

1. INTRODUCTION

Wireless sensor networks (WSN) is a network of spatially distributed physical sensors which sense the data like temperature, moisture, pressure etc. This ability of sensors is used in wide range of applications like forest fire detection, medical monitoring pollution monitoring and agriculture surveillance (Bruckner, Picus, Velik et al., 2012). There is a long history of using sensors in medicine and public health. Due the recent evolutions in electronics, it is highly economical to get the low cost, lightweight multipurpose sensors which can be integrated to use in different applications. Wireless sensors have limitations such as Non rechargeable battery power, Low computation ability, Limited memory storage and No standard platform (El Emary, & Al-Gamdi, 2014; Hamid, Harouna, Salele et al., 2013).

WSN consists of battery powered small devices which are capable of acquiring physical information from environment, processing of information and forwarding the collected information to the required destination. Application which involve multimedia data often act as sensor-actor network in which scalar sensors collects the environmental information and depending upon the environmental parameters the multimedia nodes are switched On or OFF. The multimedia nodes can be operated in sleep/awake mode as per requirement. This way is efficient way in terms of energy saving and lifetime improvement of the network (Ekleitis, Meger, Dudek, 2006).

Figure 1. Wireless Camera Sensor Network (WCSN)

Figure 1 shows architecture of Wireless Camera Sensor Network (WCSN). WCSN is constructed by a set of small and low cost sensor nodes which can produce images or videos from the sensing area. The images are then processed and transmitted to other sensors or to a central base station. The base station and the sink node in the architecture shown are static and the nodes in sensor field can be static or mobile as per application.

WCSN are widely used in various fields such as battlefield visual monitoring, environment monitoring, safety monitoring, person locator services, traffic monitoring, intelligent home, and medical treatment and public healthcare.

Pure static networks face problems due to dynamic changes of events and unpredictable environmental change in the network. Some of the challenges for static network are (Rezazadeh, 2012):

• • Initial deployment of the network would not give full coverage and connectivity of the network. The randomly deployed networks may get divided into small non-connected networks. This may degrade the performance of the application.
• • As sensor node are battery operated those may deplete their energy after some time and energy holes may be created in the network which will result in unconnected sensor field.
• • More no of node are required in case of object tracking kind of applications. No. of nodes can be reduced if some nodes are capable of relocating themselves as per network and application requirement.
• • Addition and movement of some specialized node will increase the scope of the application. For some applications it is feasible to add or remove nodes as per time schedules and tracking requirements from the network.

In military applications the scenario can be developed in which the mobile nodes act as sink node or data collector node, which collects the information from other nodes which are stationary nodes. In such cases sparsely deployed node will also be able to communicate. In smart city applications not only sink node but the sensor nodes will also be moving. In target tracking scenario combination of static and mobile nodes can be used effectively. Applications explain above consist of the movement of node which is unpredictable. The traditional algorithms for deployment, duty cycle scheduling and routing needs improvement.

Following are type of mobility which can be applied to the applications described previously.

• • Sink Node Mobility

Sink node mobility allows sink node to move between restricted area for collection of data from the static node. This avoids the creation of energy hole as well as promotes the energy balancing in the network.

• • Sensor Node Mobility

In some network mobile agents are deployed to balance the energy consumption in the network. These agents change their places as per the dynamic requirement of the network and transmit data efficiently to the sink node.

Data collection process can be improved with the help of MEs (Mobile Elements). In this type of network, the three elements are used Regular sensor nodes, Sink nodes and Special support mobile nodes. The task of the regular node is to sense the environmental phenomenon, record the information and send it to required destination. The task of the sink node is to collect information from regular and mobile node. The task of the special support mobile node is to act as intermediate collector node between sink node and static node or act as mobile gateways (Di Francesco, Das, & Anastasi, 2011).

Mobility can be of two types. In first type the nodes are moving by pre specified path or algorithm for improving efficiency of the network i.e. it is a result of deliberate movement of object or person. In other type, the node movement is due to external forces such as wind, water or disaster (Bharti, Behniwal, Sharma, 2013). In the applications such as health monitoring system and disaster recovery system the observation recording with the help of scalar node and multimedia node is required in such cases the mobility is required. The existing algorithms or methodologies which are designed for static nodes are not suitable to achieve efficient results in mobile environment.

In this paper we concentrate on the applications where sensor field is divided into clusters. Each cluster has a set of nodes which are mobile or stationary.

2. MOBILITY ISSUES

Implementing mobility in the network expands the scope of applications for the WSN but due to high mobility various issues are (Bharti, Behniwal, & Sharma, 2013):

• • Deterioration in the quality of link which in turn increases packet loss and retransmission overhead in the network.
• • Variation in routing load will cause frequent root updation.
• • Frequent route changes increases packet delivery delay in the network.
• • Location recognition before every communication is required.
• • Neighbor discovery process may add unnecessary delay in communication establishment.
• • Speed of mobility, direction of movement, pause time of node and mobility model are the critical factors while implementing MAC layer and network layer protocols.
• • For schedule based protocols frequent schedule updation will be required.
• • There may be variation in throughput of the network.

2.1. Mobility Models

Choice of the mobility model plays important role in implementation of sensor network protocol and architecture. Mobility models can be applied to group of the nodes or to the individual node movement. Basic types of mobility models and its functions are described below (Premi & Shaji, 2011; Vasanthi, Romenkumar, Ajithsingh et al., 2011; Bai & Helmy, 2004):

2.1.1. Random Based Mobility Models

• • Random Waypoint Model: In random waypoint model the main parameters are speed and direction which is randomly selected by the node. Minimum and maximum ranges for speed and direction are specified for the simulation area. Algorithm selects random values between 0 and max value after a fixed time interval called as pause time for speed and direction. The pause time for all nodes is same.
• • Random Walk Model: Random walk model is used to emulate the unpredictable movements of the objects. It is also known as Brownian motion. It is a memory less mobility process in which previous status is not referred for the future movements. Simulation has two parameters speed and mobility for which minimum and maximum values are specified. Node moves with a fixed speed and direction for specific period. And after specific period it selects other random values and proceeds further.

2.1.2. Geographic based Models

Some applications required pre specified paths and restriction for the movement of the node. Such models with restricted mobility are called Geographic based models.

• • Pathway Mobility Model: This model restricts the movement of the node to the fixed pathway on the map or simulation area. This pathway may be randomly generated or carefully defined. This model is useful for the applications in which the sensor field is known and restricted. E.g. a city, a building or the area under surveillance.
• • Obstacle Mobility Model: Another important key point in the mobility of the node is obstacles. Obstacles in the path cause the reduction in signal strength or break in communication. The movement of a node is necessary to avoid any data loss or connectivity breaks. As obstacles are coming into path the change in the must be done.
• • Manhattan Grid Model: In this model the nodes are allowed to move vertical and horizontal along the grids. This model is specifically designed for the streets which are organized in grid fashion. At each intersection of the grid the node can move to any direction with certain probability. In vehicular network this model is more effective.
• • Freeway Model: Freeway model allows movement of nodes freely along the streets with lane discipline. Main idea behind model is to collect traffic information and vehicle tracking. This model uses maps for direction and support. Safe distance is maintained between nodes to increase efficiency of the method.

2.1.3. Temporal Dependency Mobility Model

The current velocity of node may depend on previous velocity of the node. Thus velocities of nodes at different times are correlated. This correlation is called as temporal dependency.

• • Gauss-Markov Mobility Model: In this model it is assumed that velocity of node is correlated over time. Velocity is calculated as Gauss-Markov stochastic process. When node reaches boundary it is shifted by 180⁰ to keep node away from the boundary.
• • Smooth Random Mobility Model: Random movement of the node may result in unrealistic behavior of the node. This model supports incremental and smooth change in the velocity and direction of the node. In real time scenarios the node speed is always a preferred speed rather than the random speed selected by the algorithm. Hence in smooth random mobility model the probability of preferred speed is high. According to the probability distribution function the value of next velocity is decided then the value of velocity is changed incrementally from current velocity to new velocity. However, the change in the movement direction is also gradual and smooth. The change in direction is done by changing the direction values by small amount of degree.

2.1.4. Spatial Dependency Mobility Models

In other random models the velocity and direction of the node is independent of other nodes. In realistic applications this is not feasible. In vehicular ad hoc networks speed of vehicle is relative to the speed of vehicle which is ahead of it. In many scenarios the nodes follow the behavior of the leader node or the reference node. This characteristic is called special dependency.

• • Reference Point Group Mobility Model: In real time scenarios nodes while moving in the sensor field coordinate with each other and move together. In this model every group has a leader at center and this leader coordinates the movement of the member nodes in the group.

For group leader a motion vector is initialized randomly or by using pre specified paths. Each member in the group uses a mobility factor using this motion vector. Each member deviates by some degree from the value of motion vector and follows the group leader. For each member the movement of the leader node is a reference point for movement direction. The RPGM model is able to represent variety of mobility scenarios in real time.

• • A Set of Spatially Correlated Models: In (M. Sanchez and P. Manzoni, 1999) authors proposed set of mobility models which uses cooperative behavior of nodes for mobility. These models are using strong special dependency of the neighboring nodes. The examples of models are Column Mobility Model, Pursue Mobility Model and Nomadic Mobility Model.

The Column Mobility Model represents movement of set of nodes or robots to a fixed particular direction. This model is useful in Military and security applications.

The Pursue Mobility Model is used when set of nodes are trying to track / follow a particular node/external element. This Model is useful in target tracking scenario.

The Nomadic Mobility Model is used for moving set of node together to any randomly chosen direction. This model is useful in mobile communication on military fields.

3. RELATED WORKS

In introduction section various issues related to mobility are discussed. By analyzing the various mobility models and their applications we can use these mobility patterns to achieve effective results and to get benefits of the mobility in the network for various key points described below. There are varieties of advantages of applying mobility in wireless sensor network. Some of the major key points are:

3.1. Solution to Energy Hole Problem

In (Marta & Cardei, 2008) author’s uses sink mobility to overcome the problem of energy holes. In WSN the sensor nodes close to the sink tends to deplete their energy faster than other nodes. This causes the formation of energy holes. The author proposes the method to use mobile sink which will change its location according to the energy level of neighboring nodes and always tends to find the richer energy areas. The Network Lifetime is main metric considered here.

• • Sink Mobility with Pre-Established Path: In this case where the location of the sink varies at the corners of hexagon where the sink gathers the data at each corner, there is noticeable improvement in the network lifetime compared to static sink. Further if sink varies the location more often than just corners, network lifetime seems to be increased more than previous case.
• • Sink Mobility with Unrestricted Path: This case proposes the distributed algorithm. The sink moves in autonomous path. The data gathering is done in rounds using clustering method. The next movement of sink is determined by localised algorithm which considers the energy of the immediate neighbour of the sink node. Also the best location of sink is determined inside the cluster. A new algorithm is invoked to check the connectivity of sink.

For simulations a special java framework is used here. The network lifetime is measured in terms of number of rounds until the first sensor node dies. The simulation results show that proposed algorithm performs better than the static sink case while considering the network lifetime.

3.2. Target Tracking in Visual Sensor Network

In the scenario where multiple targets are considered, the sink moves to the point which has average shortest distance from the targets. The simulation results show that the mobile sink with single target prolongs the network lifetime than it in the static sink. The node scheduling also helps to increase the network lifetime. A simulation proves that mobile sinks with node scheduling helps to increase the network lifetime when the number of nodes is large. The speed of the sink also affects the network lifetime. In case of the multiple target situations, the constant speed of mobile sink gives best performance than greater speed.

The proposed scheme uses mobile sink with unlimited energy and the cluster head predetermined located at the center. The cluster heads determine the shortest path to sink by comparing hop count from sink and residual energy of nodes. The cluster head stores the routing information which is updated each time the sink moves to new location. The node selection is done by using sensing range of node inside the cluster. The movement of the sink is determined by the tracking error value lower is the tracking error for the position higher is the probability of sink to move towards the location.

The performance metric used here is network lifetime which is defined as the time from the network start time to the time when the first cluster head dies. The simulation is done in the MATLAB. Here three scenarios considered are 1. static sink 2. mobile sink with single target 3. mobile sink with multiple target. The simulation shows that the mobile sink increases the network lifetime than the static sink.

3.3. Efficient and Fair Sensor Data Collection

• • Efficient data collection (Matsuo, Goto, Kanzaki et al., 2013):

Efficiency can be achieved through data gathering based on Status of Connection with Multiple Mobile Sinks in Wireless Sensor Networks. In (Matsuo, Goto, Kanzaki et al., 2013) author proposed method to transmit sensor data acquired by sensor nodes to at least one mobile sink. The DFBRR (Data Forwarding based on Beacon Reception Rate) achieves efficient data gathering assuming the existence of a large number of mobile sinks that freely move around the target region. In DFBRR, each sensor node determines its status of connection with mobile sinks (called connectivity in proposed method) based on the beacon reception rate, which is the number of received beacons from mobile sinks in a certain period of time. In addition, information on connectivity is shared among neighboring sensor nodes. When a sensor node cannot connect directly with any mobile sinks, it forwards sensor data to some neighboring nodes based on the shared information. By doing so, DFBRR achieves efficient data gathering from sensor nodes to mobile sinks.

For implementation authors considered multiple mobile sinks and sensor nodes. Method improves data gathering ratio while suppressing traffic. An adaptive strategy can be added which adaptively changes the information sharing mechanism according to the change in mobility of mobile sinks, the frequency of sensing and so on.

Only sink mobility is not the aim of designing various protocols for mobile sensor network. Other nodes in the network can be mobile for various applications. In (Seino, Yoshihisa, Hara et al., 2010) for efficient data gathering mobile agents are used. In dense mobile sensor networks, to reduce traffic for gathering sensor data, it is desirable to efficiently gather sensor data from the minimum number of mobile sensor nodes necessary to guarantee the sensing granularity, which the application requires. In (Ekleitis, Meger, Dudek, 2006) author proposed DGUMA (Data Gathering method Using Mobile Agents) method, mobile agents aggregate multiple readings with the same value into one reading and forward the aggregated data to reduce the traffic. Moreover, extended method dynamically controls communication routes from each agent to the sink so that the agents can effectively aggregate the readings. It assumes dense MWSNs (Mobile Wireless Sensor Network) constructed by mobile sensor nodes which are held by ordinary people and equipped with a radio communication facility. The results of the simulation experiments show that DGUMA/DA (DGUMA with Data Aggregation) can gather sensor data with high delivery ratio and small traffic.

• • Fair data collection (Chen, Lai, & Wang, 2011):

Fair Sensor Data Collection (FSDC) a communication protocol to increase the data amount collected mobile sink considering fairness of communication time.

The mobile sensor collects the data from each sensor while travelling. Fairness should be considered while collecting the data. If the mobile sink does not consider fairness, it can collect a few data from a sensor even if the sensor has a large amount of sensor data. For this purpose, the author proposes the protocol which sets the upper limit to the first communication time, after which the sensor can again send the data to sink if sink is not busy with another sensor.

The proposed algorithm has following phases. Phase 1: The mobile sink first sends a beacon to search a sensor which is available for communication. After it receives the response from the sensor the communication begins which is called as first communication. Here the upper limit is given for the communication time and is said as FCT (First Communication time). After which the communication is over and again new search for sensor begins. Phase 2: If the sink receives the response from only the node it has communicated already it recommunicate with it with the new upper limit called re-communication limit RCT (Re-Communication time). The mobile sink finishes communication when the communication time reaches FCT or RCT.

The simulation is carried out using C++. The FSDC protocol gives better fairness than previously proposed protocol for the same data amount. The only drawback here is the sink must know whether it is communicating with the sensor for the first time or not. This implementation was for single sink. It can be implemented for multiple mobile sink nodes.

3.4. Increase in Lifetime of the Network

In (Cayirpunar, Urtis, & Tavli, 2013) the MGR (Mobile Multimedia Geographic Routing) protocol for MMSNs (Mobile Multimedia Sensor Network) is proposed. This protocol is evaluated for both the delays and energy. The one hop delay is the summation of the transmission delay, queuing delay, processing and propagation delay. The MGR tends to achieve the tradeoff between the energy and delay. The protocol implements strategic location concept in which the ideal location is current node’s next hop. The next hop selection is done using the strategic location strategy instead of the neighbor closest to the sink as in traditional geographical routing protocols. The MMSN proposed here can be used for using the sensor capabilities for the event description and evaluation. The MGR uses the geographic location information for routing. This enables the QOS guarantees for MMSN. The MMSN can be exploited for many intelligent applications and business applications.

In (Lv, Xu & Zheng, 2014) three base station mobility patterns are used which are random mobility, grid mobility, and spiral mobility to maximize the lifetime of wireless sensor network. To avoid the shadowing effects of specific protocols or algorithms a novel Mixed Integer Programming (MIP) framework is used which enables to explore the design space under optimal operating conditions. It takes advantage of base station mobility to maximize the network lifetime, determining the optimal mobility Pattern which is of utmost importance. It uses General Algebraic Modeling System (GAMS) for the numerical analysis of the developed MIP model.

To balance the energy consumption in the network tunable coefficients can be used (Restuccia & Das, 2015). Load balancing contributes to maximize the lifetime of the network. In (Restuccia & Das, 2015) authors proposed novel model to define lifetime of WMSN (Wireless Multimedia Sensor Network).

In (Cheng, Das, Francesco et al., 2011) swarm intelligence based algorithm is proposed which gives energy efficient solution and meets pre-defined quality of service requirements. The experiments are carried out on real tested. Results shows that algorithm is highly scalable and energy efficient.

3.5. Efficient Streaming Data Delivery

Finding efficient routing path for streaming multimedia data is a challenging task for WSNs. In (Bharti, Behniwal & Sharma, 2013) the authors making use of sink mobility for the delivery of streaming data in cluster based WSNs. Energy-efficient streaming data delivery (SDD) protocol for cluster-based WSNs with MS is used to maintain the end-to-end connectivity between the source cluster head (CH) and the mobile MS in an efficient way while avoiding the constant flooding of sink location information as the MS moves across multiple clusters. To explicitly support the sink mobility, a cross-clusters handover mechanism and a path redirection scheme is used. SDD exploits an on-demand scheme to dynamically build routes to avoid channel overhead and improve scalability. The on-demand route discovery follows the usual scheme consisting in a route request and reply message pairs. SDD is able to maintain the end-to-end connectivity between the sources and MS (Mobile Station), while avoiding the constant flooding of sink location information as the MS traverses across multiple clusters. SDD is efficient and lightweight, and thus particularly suitable for streaming of multimedia data in WSNs with MS.

4. PROPOSED METHODOLOGY

The related works section describes the details about the mobility behavior and the mobility models in detail. This section gives details about efficient methodologies for improving performance of the network using mobility.

By analyzing the benefits and drawbacks of the existing mobility model we are proposing a new mobility model MSMA - Mobile Sink with Mobile Agent based model as shown in Figure 2.

We assume the cluster based architecture for the network. The network is divided into clusters and each cluster consists of a cluster head node and set of member nodes which follows instructions from the cluster head. The member nodes of the networks are combination of scalar and multimedia node. Some of the nodes in the member nodes are mobile nodes and are called as agent nodes. These agent mobile nodes will work as an assistant to the sink node for collecting information from all member nodes and communicating it towards the sink node. Number of agent nodes in the cluster has a maximum value as 2 and minimum value as 0.

Figure 2. MSMA- Mobile Sink with Mobile Agent based model

Figure 2 shows the cluster based architecture. It consists of a mobile sink and agent nodes. Initially when network is formed the node placement is done as per the application requirement. As network progress the dynamic positions of the agent node can be change as per the network performance requirement. The functions of the Mobile sink and Mobile agent node is divided into two phases Dynamic Node Deployment phase and Efficient Data Collection phase.

• 1. Dynamic Node Deployment Phase: Initially when network is deployed, the deployment is based on the application requirements. As the network progress with the time the energy level of each node will vary. If sink node changes its position as per pre-defined path energy depletion of particular node can be avoided. As shown in Figure 2 the sink node follows fixed path as per application requirement. Some times in case if some nodes are dead due to energy depletion dynamic holes are created in the network. The position of the agent nodes can be changed as per dynamic hole positions in the network. The movement of the agent node will also help to give full coverage for all the nodes in the network. This will definitely increase the lifetime of the network.
• 2. Efficient Data Collection Phase: The Dynamic Node Deployment Phase adjusts the position of the nodes such a way that it will create a fully covered network for efficient communication. In Efficient Data Collection Phase the data collection is classified into two categories as scalar data and multimedia data. The agent nodes and sink node while collecting information from the node first analyses the type of node to which they are communicating. If the node is a multimedia node naturally the data transmission required will be more and will take more time. If the node is a scalar node the duration of data transfer will be less as compare to multimedia node. Before actual data transmission various parameters are shared and analyzed for processing.

Mobility parameters for Mobile Sink and Mobile Agent based Mobility Model:

By analyzing the related works in Section 3 and 4, following parameters can be listed for mobility management in various protocols:

• • Current location of node
• • Mobility speed.
• • Pause time of the node.
• • Range of the node.
• • Link Quality Index.
• • Remaining Energy of node.

In section 3 authors proposed variety of methodologies for using mobility to improve the performance of the network. All the current protocols and the methodologies are application specific and concentrate on particular parameter. The proposed mobility model is suitable for large set of applications. Different tunable parameter setting is possible with this model if requirement changes drastically.

Adding mobility to the sensor networks enhances capability of application for real world. The applications like monitoring daily activities of patient or monitoring activities in a room can be implemented effectively with combination of multimedia devices and mobility (Grünerbl, Bahle, Hanser et al., 2012).

5. CONCLUSION

In this paper, we have studied impact of mobility on sensor network and various methodologies proposed by authors to use mobility effectively for different applications. We have also discussed mobility issues and challenges in variety of applications. The mobility issues section describes in detail about the issue and effects of mobility for future applications. We also proposed effective mobility model in proposed methodology section for improving lifetime of the network. The Mobile Sink with Mobile Agent mobility model uses dynamic node deployment and efficient data collection for lifetime improvement of the network. Consideration of the mobility parameters in accordance with challenges leads to effective solutions for future internet or Internet of Things (IoT).

This research was previously published in the International Journal of Rough Sets and Data Analysis (IJRSDA), 4(2); edited by Nilanjan Dey; pages 24-35, copyright year 2017 by IGI Publishing (an imprint of IGI Global).

REFERENCES

Bai, F., & Helmy, A. (2004). A survey of mobility models. Wireless Adhoc Networks . USA: University of Southern California.

Bharti, D., Behniwal, M., & Sharma, A. K. (2013). Performance analysis and mobility management in wireless sensor network. International Journal of Advanced Research in Computer Science and Software Engineering , 3(7), 1333–1342.

Bruckner, D., Picus, C., Velik, R., Herzner, W., & Zucker, G. (2012). Hierarchical semantic processing architecture for smart sensors in surveillance networks. IEEE Transactions on Industrial Informatics , 8(2), 291–301.

Cayirpunar, O., Urtis, E. K., & Tavli, B. (2013, September). The impact of base station mobility patterns on Wireless Sensor Network lifetime. In PIMRC (pp. 2701-2706). doi:10.1109/PIMRC.2013.6666605

Chen, M., Lai, C. F., & Wang, H. (2011). Mobile multimedia sensor networks: Architecture and routing. EURASIP Journal on Wireless Communications and Networking , 2011(1), 1–9. doi:10.1155/2011/103027

Cheng, L., Das, S. K., Francesco, M. D., Chen, C., Ma, J., & Xie, D. (2011, June). Streaming data delivery in multi-hop cluster-based wireless sensor networks with mobile sinks. Proceedings of the 2011 IEEE International Symposium on a World of Wireless, Mobile and Multimedia Networks (WoWMoM) (pp. 1-9). IEEE. 10.1109/WoWMoM.2011.5986379

Di Francesco, M., Das, S. K., & Anastasi, G. (2011). Data collection in wireless sensor networks with mobile elements: A survey. ACM Transactions on Sensor Networks , 8(1), 7. doi:10.1145/1993042.1993049

Dingli, A., Attard, D., & Mamo, R. (2012). Turning homes into low-cost ambient assisted living environments. International Journal of Ambient Computing and Intelligence , 4(2), 1–23. doi:10.4018/jaci.2012040101

Ekleitis, I., Meger, D., & Dudek, G. (2006). Simultaneous planning, localization, and mapping in a camera sensor network. Robotics and Autonomous Systems , 54(11).

El Emary, I. M., & Al-Gamdi, A. H. (2014). Wireless Sensor Networks with its Effective Impact in the Health Care Application. Journal of Applied Medical Sciences , 3(3), 5–18.

Grünerbl, A., Bahle, G., Hanser, F., & Lukowicz, P. (2013). Uwb indoor location for monitoring dementia patients: The challenges and perception of a real-life deployment. International Journal of Ambient Computing and Intelligence , 5(4), 45–59. doi:10.4018/ijaci.2013100104

Hamid, N. I. B., Harouna, M. T., Salele, N., & Muhammad, R. (2013). Comparative Analysis of Various Wireless Multimedia Sensor Networks for Telemedicine. International Journal of Computers and Applications , 73(16).

Kanzaki, A., Yamamoto, A., Hara, T., & Nishio, S. (2012, November). Data forwarding method based on status of connection with multiple mobile sinks in wireless sensor networks. Proceedings of the 2012 Seventh International Conference on Broadband, Wireless Computing, Communication and Applications (BWCCA) (pp. 170-177). IEEE. 10.1109/BWCCA.2012.36

Lv, B., Xu, J., & Zheng, X. (2014, September). Modeling the lifetime of wireless multimedia sensor networks with a mobile sink. Proceedings of the 2014 IEEE 25th Annual International Symposium on Personal, Indoor, and Mobile Radio Communication (PIMRC) (pp. 1814-1818). IEEE. 10.1109/PIMRC.2014.7136464

Marta, M., & Cardei, M. (2008, June). Using sink mobility to increase wireless sensor networks lifetime. Proceedings of the 2008 International Symposium on a World of Wireless, Mobile and Multimedia Networks WoWMoM ‘08 (pp. 1-10). IEEE. 10.1109/WOWMOM.2008.4594857

Matsuo, K., Goto, K., Kanzaki, A., Hara, T., & Nishio, S. (2013, March). Data gathering considering geographical distribution of data values in dense mobile wireless sensor networks. Proceedings of the 2013 IEEE 27th International Conference on Advanced Information Networking and Applications (AINA) (pp. 445-452). IEEE. 10.1109/AINA.2013.11

Premi, M. G., & Shaji, K. S. (2011). Impact of mobility models on MMS routing in wireless sensor networks. International Journal of Computers and Applications , 22(9), 47–51. doi:10.5120/2608-3638

Restuccia, F., & Das, S. K. (2015, June). Lifetime optimization with QoS of sensor networks with uncontrollable mobile sinks. Proceedings of the 2015 IEEE 16th International Symposium on a World of Wireless, Mobile and Multimedia Networks (WoWMoM) (pp. 1-9). IEEE. 10.1109/WoWMoM.2015.7158130

Rezazadeh, J. (2012). Mobile wireless sensor networks overview. International Journal of Computer Communications and Networks , 2(1).

Sanchez, M., & Manzoni, P. (1999, January). A Java-Based Ad Hoc Networks Simulator. Proceedings of the SCS Western Multiconference Web-based Simulation Track.

Seino, W., Yoshihisa, T., Hara, T., & Nishio, S. (2010, November). A communication protocol to improve fairness and data amount on sensor data collection with a mobile sink. Proceedings of the 2010 International Conference on Broadband, Wireless Computing, Communication and Applications (BWCCA) (pp. 33-40). IEEE. 10.1109/BWCCA.2010.45

Vasanthi, V., Romenkumar, M., Ajithsingh, N., & Hemalatha, M. (2011). A detailed study of mobility models in wireless sensor networks. Journal of Theoretical and Applied Information Technology , 33(1), 7–14.