Impact of the Quality of Spatial 3D City Models on Sensor Networks Placement Optimization

Abstract

<jats:p> Sensor networks are increasingly used for tracking, monitoring, and observing spatial dynamic phenomena in the real world (e.g. urban area). In order to ensure an efficient deployment of a sensor network, several optimization algorithms have been proposed in recent years. Most of these algorithms often rely on oversimplified sensor models. In addition, they do not consider information on the terrain topography, city models, and the presence of diverse obstacles in the sensing area (e.g. buildings, trees, poles). Only some of those optimization algorithms attempt to consider the terrain information in the optimization of a sensor network deployment. However, most of these algorithms consider that the spatial models used for this purpose are perfect representations of the reality and are not sensitive to the quality of the information. However, spatial models are simplified representations of a complex reality, and hence are inherently uncertain. In this paper we will investigate the impact of the spatial data quality on the optimization of a sensor network and its spatial coverage in an urban area. For this purpose, we will investigate specific implications of spatial data quality criteria for a 3D city model that will be used in sensor network optimization algorithms. Then, we will analyze the impact of some of those criteria on the estimation of sensor network coverage.!Afterwards, a case study for sensor network deployment in an urban area will be presented. This case study will demonstrate the impact of 3D city models quality on the estimation of coverage using global and local optimization algorithms. Finally, the results obtained from this experimentation will be presented and discussed. </jats:p>

HybVOR: A Voronoi-Based 3D GIS Approach for Camera Surveillance Network Placement

“…The main objective of this paper is to propose a novel method that belongs to the landscape based approach. Our proposed method is also based on the space segmentation using Voronoi Diagram (as the works of [19][20][21]). However, as detailed in the following sections, the proposed solution differs from other methods found in the literature in two fundamental aspects: (1) the choice of objects that generates the Voronoi Diagram and (2) The method of assessing the level of coverage.…”

“…On one hand, Voronoi-based methods that are presented in previous work ( [19][20][21]) use the expected position of the sensors (cameras) to generate the Voronoi diagram. Each Voronoi cell will contain the objects of the environment that may be monitored by the corresponding sensor.…”

“…A Voronoi based algorithm was developed in [19] that integrates a Digital Surface Model of the area of interest. This Algorithm was inspired from the work of [20] which proposes three Voronoi strategies: (1) Vector-Based (VEC), (2) Voronoi-Based (VOR) and (3) Minimax.…”

“…This Algorithm was inspired from the work of [20] which proposes three Voronoi strategies: (1) Vector-Based (VEC), (2) Voronoi-Based (VOR) and (3) Minimax. The authors in [19] propose first to generate a Voronoi Diagram from an initial position of sensors. Then, the sensor network is optimized by moving each sensor towards its farthest Voronoi vertex until reaching the highest elevation inside the Voronoi Cell.…”

Coevrage Estimation of Geosensor in 3d Vector Environments

“…On the other hand, in the area-based coverage problem, which is the concern of this paper, the objective is to obtain the maximum region covered by geosensors, and is usually evaluated as the ratio of the covered area to the whole area (Huang and Tseng, 2005). The area-based coverage calculation methods are classified into: (i) the methods that consider a raster environment (Akbarzadeh et al, 2013;Argany et al, 2012;Cortés et al, 2004), which are limited by the spatial resolution; and (ii) the methods that model the environment as a vector dataset (Ghosh and Das, 2006;Guvensan and Yavuz, 2011;Ma et al, 2009;Cao, 2006, 2011), which have been mostly proposed for 2D spaces and do not consider the earth topography and human-made obstacles. Furthermore, the sensing models (i.e., binary or probabilistic, omnidirectional or directional, etc.)…”