Article Preview
Top1. Introduction
Wireless sensor networks (WSNs) consist of a large number of low power tiny sensor nodes (SNs) that sense data from a variety of ambient sources such as light, sound, heat, motion and transmit the sensed data to the sink. In WSN clustering, SNs are assigned to their respective cluster heads (CHs) which aggregates data emanating from each SN and finally forwards it to the sink (Fellah, & Kaddour, 2017; Moedjiono & Kusdaryono, 2013; Ahmad et. al., 2013). CHs are classified into relay and normal CHs. Relay CHs are those which are very close to the sink and transmit data directly to the sink. Whereas normal CHs are those which are far away from the sink and transmit data to the sink through relay CHs (Kim et. al., 2010; Liang et. al., 2010; Nazir & Hasbullah, 2010; Mottaghi & Zahabi, 2014). Hence, life of relay CHs becomes low due to excessive data relay load and it will die early, resulting in energy holes in the network. One way to solve the energy hole problem is to use mobile sink (MS) as it minimizes the load on CHs by moving closer to CHs for data aggregation. Use of MS as a solution to the energy hole problem was widely adopted in the last decade. Most algorithms in the past have focused on dividing the deployment region into sub regions and using MS for data aggregation. But MS can also lead to energy holes due to its dynamic nature and random positions inside the sub regions, resulting in excessive data relay load on relay CH near the MS as shown in Figure 1. To avoid this, we propose to distribute the data relay load optimally to multiple CHs and locating MS near these multiple CHs.
Figure 1. Excessive data relay load on relay CH near the sink
Furthermore, routing of sensed data to MS inside the sub regions of a MS WSN is important for energy conservation in a WSN. Proposed algorithm optimally selects the routing CHs as per data transfer requirements of SNs and CHs to minimize energy consumed by sensor nodes for transmitting data from sensor nodes to MS through CHs. Identifying optimum location of the MS to distribute data relay load from one CH to multiple CHs and obtaining optimum routing combination of normal CHs is a combinatorial optimization problem that can be solved by meta heuristic algorithms. Recently, algorithms like ant bee colony (ABC) and artificial neural networks are used prolifically for optimization of WSN parameters like sensor lifetime, load balancing and residual energy of sensors. Biogeography Based optimization (BBO) (Simon, 2008) is a meta-heuristic algorithm that imitates the behaviour of biogeography to solve real-world optimization problems. BBO has not only shown better results than other metaheuristic algorithms like ABC and PSO but has also been used in many applications such as remote sensing, intelligent battlefield preparation, and traveling tournament problem. This motivated us to use BBO to optimize the performance parameters of a MS WSN. Author’s contribution to the paper can be summarized below:
- •
We propose an adaptive MS algorithm that distributes the data relay load optimally to multiple relay CHs and locates the MS near these multiple CHs;
- •
The proposed MSDR-BBO algorithm optimally selects the routing CHs as per data transfer requirements of SNs and CHs to minimize energy consumed by sensor nodes for transmitting data from sensor nodes to MS through CHs;
- •
The proposed MSDR-BBO algorithm ensures full and hole free coverage of the deployment region using MS in minimum time.
MSDR-BBO algorithm is validated using Matlab and Netsim and it performs better than algorithms like Nested routing (Yarinezhad, 2019), SENMA (Ang et. al., 2018) and CMS2TO (Gharaei et. al., 2018). The literature review and advantages of mobile sink are presented in section 2 and section 3 respectively. Proposed MSDR-BBO algorithm is explained in section 4. Results are explained in section 5 and we conclude in section 6.