Hole Detection for Quantifying Connectivity in Wireless Sensor Networks: A Survey

Table 2

Strengths and shortcomings of coverage hole detection algorithms.

Researcher	Year	Concept used	Advantages	Disadvantages
Wang et al. [33]	2003	Using Voronoi diagram	Each node needs to monitor small area around it; low communication complexity	Performance depends on ratio of mobile to static sensors; no obstacle modeling
Guiling et al. [26]	2004	Using computational geometry	Extensible to large deployments because communication and movements are local	Poor performance on initial clustered deployment and lower communication range
Fekete et al. [21]	2004	Average density and centrality density	Avoids GPS	Complex computations; unrealistic assumption on sensor distribution and density
Ghrist and Muhammad [25]	2005	Using homology	Works in coordinate free sensor networks	Not scalable; complex calculations
Bldg and Funke [28]	2005	Using isosets and isocontours	Works for both continuous and discrete cases and avoids GPS	Requires high network density and huge overhead in selecting seed nodes and constructing isolevels
Kröller et al. [30]	2006	Using combinatorial structures	Supports quasi-UDG	Fails in sparse n/w and deals with complex combinatorial structures
Fang et al. [23]	2006	Using stuck nodes	No assumption about node distribution	High message complexity and UDG constraint
Wang et al. [27]	2006	Using shortest path tree	No UDG constraint; works for random distribution	Needs high node density, high communication overhead due to flooding and cannot detect multiple adjacent holes
Bi et al. [16]	2006	Cooperative neighbor and coverage ratio	Scalable	Not suitable for randomly deployed dense networks
Corke et al. [22]	2007	Using convex hulls	Can discover holes based on normal message traffic	Local detection algorithm needs initial state information; global detection algorithm needs location information
Li and Hunter [31]	2008	Using homology and connectivity information	Scalable, large holes can be easily detected	Fails for trivial holes
Dong et al. [20]	2009	Using FGP transformation	Can detect small holes accurately; correctness of hole detection is proved	Communication overhead
Yang and Fei [14]	2010	Using TENT rule	Single node can detect hole efficiently	Needs localized nodes
Zhao et al. [32]	2011	Mathematical analysis	Can find precise location of holes	Needs location information and works only for mobile nodes
Martins et al. [17]	2011	Rips and Cech complexes	Can detect nontriangular holes efficiently	Communication overhead
Zeadally et al. [18]	2012	Using hop based approach	Works for node degree 7 or higher	Communication overhead to identify x-hop neighbors
Babaie and Pirahesh [29]	2012	Using triangular oriented structure	Exact hole area is identified	Needs localized nodes
Chu and Ssu [19]	2014	Using contour lines	No UDG constraint; suitable for dynamic hole detection	Communication overhead
Senouci et al. [24]	2014	Using TENT rule and virtual forces’ concept	Minimizes resource consumption	Cannot detect holes at the network boundaries; UDG constraint