Quasi-optimal scheduling algorithm for area coverage in multi-functional sensor networks

Abstract

Wireless Sensor Networks (WSNs) are widely used to perform measurement and monitoring tasks in large sensing fields. Several applications require every point in the monitored area to be covered by at least k sensors (k-coverage problem) for measurement accuracy or robustness. In heterogeneous multi-functional sensor networks, p parallel observation tasks are performed with p types of sensors, each task having its own ki(i = 1, 2, . . . , p) coverage requirement, assuming that multi-functional sensors are used which may have multiple measurement capabilities (multi-k-coverage problem). In order to prolong the lifetime of the network, active and sleeping states of sensors can be altered while maintaining the required coverage for all sensing tasks. In this paper an efficient distributed sensor state scheduling algorithm for multi-functional WSNs is proposed to solve the multi-k-coverage problem. The proposed multi-functional controlled greedy sleep algorithm is the generalisation of a recently proposed quasi-optimal scheduling algorithm. It is easy to implement and solves the common scheduling problem of heterogeneous multi-functional sensor networks. It has low local communication overhead and can adapt to dynamic changes in the network, while the required network-wide coverage for all sensing tasks is guaranteed as long as it is physically possible. The performance and the fault tolerance of the algorithm are illustrated by simulation examples.

An Energy-Efficient Multiobjective Scheduling Model for Monitoring in Internet of Things

“…The total energy consumption of a node must be at most equal to its initial energy. Several propositions were formulated in the literature to address MLP and problem variations [5], [17], [24], [25].…”

“…This is due to the fact that the input of Phase III is only the unique VCs assigned to monitor the periods (u) from Phase II. TSP path is solved using Held-Karp dynamic programming algorithm [17]. There are at most (2 u .u) subproblems, each of which takes linear time (u) to solve.…”

“…A prevalent approach toward expanding the life span of such networks is by using sleep scheduling. In sleep scheduling, nodes enter sleep state frequently, and intermittently wake up to check for action in the network [16], [17]. For critical and time-sensitive IoT applications, monitoring needs to be continuous regardless of the lack of activity in the network.…”