Detecting Coverage Holes in Wireless Sensor Networks Using Connected Component Labeling and Force-Directed Algorithms

TL;DR

This work tackles detecting coverage holes in wireless sensor networks without relying on node coordinates or sensing ranges. It introduces FD-CCL, a coordinate-free pipeline that first generates a graph layout using Force-Directed algorithms and then applies Connected Component Labeling to identify and characterize holes, contrasting it with Contour Tracing-based FD approaches. The authors demonstrate that FD-CCL achieves comparable accuracy to FD-CT while offering substantially faster hole locating and richer hole properties through labeling, with improvements most pronounced on larger, more connected graphs. The approach holds practical promise for energy-efficient WSN monitoring and rapid hole localization in security and surveillance contexts.

Abstract

Contour detection in Wireless Sensor Networks (WSNs) is crucial for tasks like energy saving and network optimization, especially in security and surveillance applications. Coverage holes, where data transmission is not achievable, are a significant issue caused by factors such as energy depletion and physical damage. Traditional methods for detecting these holes often suffer from inaccuracy, low processing speed, and high energy consumption, relying heavily on physical information like node coordinates and sensing range. To address these challenges, we propose a novel, coordinate-free coverage hole detection method using Connected Component Labeling (CCL) and Force-Directed (FD) algorithms, termed FD-CCL. This method does not require node coordinates or sensing range information. We also investigate Suzuki's Contour Tracing (CT) algorithm and compare its performance with CCL on various FD graphs. Our experiments demonstrate the effectiveness of FD-CCL in terms of processing time and accuracy. Simulation results confirm the superiority of FD-CCL in detecting and locating coverage holes in WSNs.

Figures (15)

• Figure 1:
• Figure 2: Steps in Connected Component Labeling process.
• Figure 3: The main procedure of FD-CCL.
• Figure 4: Examples of Ground Truth Dataset.
• Figure 5: Examples of sparse and uniform layouts generated by force-directed algorithms.