15.4 standard [12] of the ns-2 framework [13]. Finally, thanks to the insight provided by the simulation results, a real video surveillance and sensing monitoring application has also been implemented and intensively tested. To do so, we have programmed hardware prototypes, based on the Imote2 wireless module [14,15] such for video capture, and MicaZ devices [16] for sensing monitoring of physical parameters, which constitute the nodes of a WMSN deployed in an agriculture environment.The rest of the paper is organized as follows: Section 2 summarizes the related work found in the open literature. In Section 3 the optimization problem is formulated and solved by means of the goal programming multi-objective technique, and the results obtained are analyzed. Our load balancing algorithm LOAM is presented in Section 4.
Section 5 shows and comparatively discusses the performance evaluation results obtained from analysis and simulation. Section 6 describes the details of the real implementation and presents the experimental results measured, which further validate the former values. Finally, Section 7 concludes this paper.2.?Related WorkDue to the strict limitations of power supply, memory storage and processing capacity of the WSN devices, there is a large amount of scientific literature devoted to optimizing different metrics such as lifetime, latency or reliability [1,17]. However, most of these metrics are in conflict with each other, what leads to the need to solve complex problems. For this reason, most of the works reviewed simplify the problem formulation, just optimizing a single metric (e.
g., [18,19]) or conducting a process where the selected metrics are optimized sequentially (not simultaneously). In order to do so, linear/non-linear programming techniques are used [2,3,6,20]. As a consequence, the solutions provided are not appropriate because the fully optimization of a metric does not imply optimal results for the other performance figures.In this context, Hou et al. [2] presented a solution that fairly balances the rate allocation in hierarchical (cluster) topologies. To achieve it, the authors employ linear programming and polynomial-time algorithms to firstly maximize the information that each cluster-head can collect. This result is then introduced as a constraint into a second optimization step, aimed at maximizing the traffic load of all nodes until one or more nodes reach their Carfilzomib energy-limited capacity for a given network lifetime requirement.
In [3], a cross-layer architecture for WMSN is used to minimize the end-to-end delay and to maximize the total data gathered at the
Context recognition is a highly active research area due to its large number of potential applications such as in healthcare, virtual reality, Nilotinib Leukemia security, surveillance, and advanced user interface systems.