Research On Coverage Control In Wireless Sensor Networks

Wireless sensor networks (WSNs) are integrated networks which can perform information gathering, processing and delivering. There are wide applications for WSN in industry, agriculture, military affairs, environment monitoring, biomedicine, city managing and disaster succoring. Coverage control is a basic issue of quality of service (QoS) on WSNs. The goal of coverage control is to distribute sensors to sense the monitored space and targeted object, so that WSNs can gather complete and valid information. Coverage control determines the monitoring performance on physical world for WSNs, so it is the only road to accelerate the practicability of WSNs. However, WSNs are energy-constrained, large scale and dynamic networks, which bring us great challenges on coverage research. Furthermore, WSNs are application-driven networks, and new applications are emerging in endlessly. Therefore, many new problems on coverage control are appearing and urgent to be solved.The sensing model of a sensor directly determines the coverage range and monitoring capability, so it is the research basis of coverage control. Since the information gathering of sensors is application-driven, different types of sensors have different sensing capability. Most existing work is based on disk sensing model, so the diversity of sensing models is not supported. At the same time, most work studies the coverage issue on stationary sensors. However, with the development of mobile sensors, mobile sensors have more advantages in new applications. This paper focuses on coverage issues on new sensing models. Furthermore, this paper explores new application scenarios and studies the coverage with mobile sensors. We study these important problems of stationary and mobile networks based on variant sensing models.In stationary networks, this paper focuses on directional sensing models and studies how to improve coverage in randomly deployed sensor networks. We prove that scheduling working directions of sensors to maximize directional covered area (MDAC) is NP complete. Then we propose a distributed greedy algorithm (DGreedy) to solve MDAC problem. In DGreedy algorithm, every sensor chooses the least overlapped direction, i.e., it works at the direction where it can cover maximal additional area. Further, we design a probability enhanced greedy algorithm (PGreedy). In PGreedy algorithm, the sensor which has the maximal contribution decides its work direction first. The simulations show that PGreedy algorithm remarkably improves the covered area of WSNs, so it reduces coverage holes and improves the network QoS.Except for directional sensing model, this paper further studies the sensing capability of practical sensors and proposes a new sensing model, point sensing model, which enriches the sensing models. In this paper, we use point sensing sensors to monitor diffusible events. An event can be detected only if its diffused area reaches the location of a sensor. This paper analyses the network capability about monitored event radius, based on Voronoi diagram and Delaunay triangulation. Furthermore, we design a centralized algorithm and a distributed algorithm respectively to determine the network monitoring capability. Finally, we sort the event detection time of sensors and judge their distances sequence away from the event cradleland. Then we propose a practical algorithm to determine the event cradleland by plane partition. The proposed algorithms are testified with theory analyses and simulations.In mobile networks, this paper further studying periodical target monitoring based on point sensing model. First, we prove that minimal mobile sensors sweep coverage problem (MSSC) is NP hard. Then we assume that all targets have the same sweep period for MSSC problem, and propose a centralized algorithm CSWEEP, with approximation ratio 2+e, for this simplified MSSC problem. Furthermore, we extend to the general case of MSSC problem and design a centralized algorithm GSWEEP with approximation 3. For practicality and scalability, we further designing a distributed algorithm DSWEEP. In DSWEEP algorithm, every individual sensor decides its moving path on runtime according to knowing the trajectories of other sensors. The simulations show that DSWEEP algorithm can utilize much fewer sensors to monitor a large number of targets, so that it significantly reduces the energy consumption while guaranteeing the network coverage capability.In order to further making use of the advantages of mobile sensors, this paper studies the deployment of mobile sensors by taking example for disk sensing model. In randomly deployed sensor networks, a large number of sensors are required to be deployed to provide barrier coverage, which leads to significant waste of sensors. In order to reduce the waste, this paper utilizes the relocation capability of mobile sensors. First, we define the relocation problem of minimal energy consumption barrier coverage (MEBC). Then we propose a centralized barrier coverage algorithm (CBarrier), which minimizes the total moving distances of all sensors. Considering large scale networks, this paper proposes a virtual force model and further designs a distributed algorithm DBarrier, which utilizes virtual forces to relocate mobile sensors. The simulations show that the proposed algorithms can significantly reduce the required sensors, and at the same time improve the flexibility and reliability of deployment.In summary, aiming at enriching sensing models and improving coverage capability, this paper studies new problems about new sensing models and supporting mobile sensors. This work has academic and practical value for advancing the theory and practicability of coverage control in wireless sensor networks.