**3. Related work**

According to Akkaya, Younis and Bangad [9],, finding an optimal location for the sink in a multi-hop network is a complex problem, NP hard in nature. The complexity results mainly from two factors. The first factor is the potentially infinite possible positions that the gateway can be moved to. Second, for every interim solution considered during the search for an optimal location, a new multi-hop network topology needs to be established in order to qualify that interim solution in comparison to the current or previously picked location in the search. A mathematical formulation of the problem would involve a huge number of parameters including the positions of all deployed sensors, their state information such as energy level, transmission range, etc., and the sources of data in the networks.. The authors propose moving the sink to the top relay nodes location. The sink is assumed to know the geographical location of deployed sensors. In the solution proposed in this article, an optimum location is not sought but an optimum path for a mobile sink that will ensure equitable usage of all nodes to transport data messages to the mobile sink node.

Research undertaken by Somasundara et. al [10] shows that the energy consumption in a network using a mobile base station is significantly less than that of a static network. The authors propose moving the base station around the application area. When the base station is within range of sensor nodes, it collects event data. This is not an optimum real-time solution as the sensor nodes have to wait for the base station to arrive before transmitting event information but is feasible in delay-tolerant applications such as environmental monitoring. A key difference between this researcher's proposed ideas and the model presented here is that in the model presented here an optimum path within the application area, along which one or more mobile sinks travel is calculated.

Huang, Zhai and Fang [11] consider a wireless network where the sensors are mobile, (applications such as tracking free-ranging animals, both wild or farm livestock). The problem focused on in this paper is on improving the robustness of routing when there are path breakages in the communication channel due to node mobility. The suggested solution is the use of a cooperative, distributed routing protocol to combat path breakages. The writers assume that the intended path or route between the source and destination is already known and neighbouring nodes can be used if the communication channel on the intended path fails. Our primary research focus in this paper is the calculation of an optimum path for a mobile sink to reduce the number of messages required to be re-transmitted when sending a message to a sink in the WSN. However, we will have to take cognisance of possible path breakages that may occur during the development of optimal routes.

Vupputuri, Rachuri and Ram Murthy [12] use mobile data collectors to achieve energy efficient and reliable data communication. When an event occurs, sensor nodes inform the nearest data collector. The data collector aggregates the event information and with a specified reliability factor (R) informs the base station. The primary focus of the authors' investigation is determining a mobile strategy for the data collectors to ensure reliable and energy efficient event reporting. The mobility strategy does not consider how to optimise the changing locations of the data collectors. The authors focus on reducing the number of messages sent and received by nodes close to the base station to improve network lifetime as well as ensuring that multiple paths are used to improve network reliability.

52 Wireless Sensor Networks – Technology and Protocols

According to Akkaya, Younis and Bangad [9],, finding an optimal location for the sink in a multi-hop network is a complex problem, NP hard in nature. The complexity results mainly from two factors. The first factor is the potentially infinite possible positions that the gateway can be moved to. Second, for every interim solution considered during the search for an optimal location, a new multi-hop network topology needs to be established in order to qualify that interim solution in comparison to the current or previously picked location in the search. A mathematical formulation of the problem would involve a huge number of parameters including the positions of all deployed sensors, their state information such as energy level, transmission range, etc., and the sources of data in the networks.. The authors propose moving the sink to the top relay nodes location. The sink is assumed to know the geographical location of deployed sensors. In the solution proposed in this article, an optimum location is not sought but an optimum path for a mobile sink that will ensure

equitable usage of all nodes to transport data messages to the mobile sink node.

area, along which one or more mobile sinks travel is calculated.

that may occur during the development of optimal routes.

Research undertaken by Somasundara et. al [10] shows that the energy consumption in a network using a mobile base station is significantly less than that of a static network. The authors propose moving the base station around the application area. When the base station is within range of sensor nodes, it collects event data. This is not an optimum real-time solution as the sensor nodes have to wait for the base station to arrive before transmitting event information but is feasible in delay-tolerant applications such as environmental monitoring. A key difference between this researcher's proposed ideas and the model presented here is that in the model presented here an optimum path within the application

Huang, Zhai and Fang [11] consider a wireless network where the sensors are mobile, (applications such as tracking free-ranging animals, both wild or farm livestock). The problem focused on in this paper is on improving the robustness of routing when there are path breakages in the communication channel due to node mobility. The suggested solution is the use of a cooperative, distributed routing protocol to combat path breakages. The writers assume that the intended path or route between the source and destination is already known and neighbouring nodes can be used if the communication channel on the intended path fails. Our primary research focus in this paper is the calculation of an optimum path for a mobile sink to reduce the number of messages required to be re-transmitted when sending a message to a sink in the WSN. However, we will have to take cognisance of possible path breakages

Vupputuri, Rachuri and Ram Murthy [12] use mobile data collectors to achieve energy efficient and reliable data communication. When an event occurs, sensor nodes inform the nearest data collector. The data collector aggregates the event information and with a specified reliability factor (R) informs the base station. The primary focus of the authors' investigation is determining a mobile strategy for the data collectors to ensure reliable and energy efficient event reporting. The mobility strategy does not consider how to optimise the changing locations of the data collectors. The authors focus on reducing the number of

**3. Related work** 

Gu, Bozdag and Brewer [13] use a partitioning-based algorithm to schedule the movements of mobile sinks in order to reduce data loss due to buffer overflow while waiting for a sink to arrive. This aspect is ignored in our proposed solution. Other recent research activity in this field, include the work of Marta and Cardei [14] where mobile sinks change their location when the nearby sensors' energy becomes low, and determines the new location by searching for zones where sensors have more energy. Heinzelman, Chandrakasan and Balakrishnan [15] have proposed have proposed a combination hierarchical and cluster based scheme that groups sensors and appoints a cluster head to transmit messages to the sink, thus saving the surrounding nodes energy (LEACH). The small percentage of cluster heads are randomly re-selected to improve node longevity of nodes located close to cluster heads. Patel, Venkatesan and Chandrasekaran [16] propose a Lexicographic Maximum Lifetime Vector routing scheme to maximise the first, second and so forth set of nodes time until their battery energies are depleted.

The use of a mobile relay to route all traffic passing through a static node for a specified period of time, is discussed by Wang et. al. The mobile relay traverses a concentric circle that stays within a two-hop radius of the sink. The authors show that the use of a mobile relay can improve a WSN's lifetime by 130%. Additional experiments show that a mobile sink, moving around the perimeter of a large and dense network, can best optimise WSN lifetime compared to a mobile relay or using resource rich static relays located close to a static sink [17]. The results of this paper indicate that the mobile relay should be a maximum of two hops from a static sink and that only nodes within a maximum of 22 hops from the sink need to be aware of the location of the mobile relay. The use of both a mobile sink and a mobile relay prevent over-utilisation of static nodes located close to the sink to route messages to the sink and hence increase overall WSN lifetime. We do not consider the use of a mobile relay in the solution discussed in this chapter and focus exclusively on an optimum path for a mobile sink to follow within a WSN application area.

A multi-sink heuristic algorithm (HOP) is proposed by Ben Saad and Tourancheau to find the best way to move mobile sinks in order to improve the lifetime of large scale sensor networks. Sinks are relocated to nodes located the maximum number of hops from a sink as it is assumed that these node will have higher residual energy as the nodes will not be required to re-transmit messages destined for a sink [18]. The minimum amount of time a sink will spend at a specific location is 30 days. The proposed algorithm is compared against schemes using static sinks, sinks moving along the periphery of the network, sinks moving randomly and sinks moving according to an Integer Linear Programming algorithm, in terms of network lifetime and residual energy at each sensor node. The results of simulations indicate the HOP algorithm achieves significant improvement in network lifetime over the other algorithms and that there is more even distribution of residual energy per sensor node. The HOP algorithm differs from the solution proposed in this paper, because HOP assumes that the sinks are not continuously mobile but are moved after a specified number of days to different locations within the building.
