Article Preview
TopIntroduction
WSNs comprise a large number of nodes, each of which integrates with one or more sensors, a processing subsystem and a radio transceiver (Akyildiz, Su, Sankarasubramaniam, & Cayirci, 2002). They are randomly deployed in a monitored and controlled area and are responsible for real time information collection, processing and transmission. Sensor nodes in WSNs are small in size and having low computational and communication capabilities. In most of the applications of these kinds of networks, nodes are deployed randomly without any central assistance (Bakhouya & Gaber, 2008; Platonov, & Bergman, 2011; Sabooret, Hörbst, & Ammenwerth, 2013).
In WSNs, the information collected by sensor nodes is required to be transmitted to the base station directly (single hop) or with the help of intermediate nodes (multi hop). For efficient data transmission, most of the applications require location information of the sender node (Jain, 2017). Moreover, location information (Stoleru, He, Mathiharan, George, & Stankovic, 2012; Targowski, 2014) is indispensable in most commonly used geography-based routing protocols and for implementing efficient MAC layer protocols. To estimate the location of a sensor node, Global Positioning System (GPS) (Nait-Sidi-Moh, Bakhouya, Gaber, & Wack, 2013) hardware can be attached to each sensor node. But the use of GPS in each sensor node makes network expensive and also a GPS equipped sensor node will result in more power consumption and is comparatively larger in size. Thus, it is not practically feasible to equip all the nodes with GPS. However, the use of few GPS equipped nodes i.e. as anchor nodes, to assist the other normal nodes in finding their location is practically viable and has been successfully used in several localization methods (Han et al., 2016).
In literature, under the constraints of WSNs, the localization of nodes has been addressed by various range based and range free localization algorithms (Yun, Lee, Chung, Kim, & Kim, 2009). The range-based algorithms are the methods that use point to point distance estimates or orientation information between neighbor nodes for localization. They provide higher location accuracy but require additional hardware for measurement of distance or angle information. However, range free algorithms do not need absolute distance or orientation information between the nodes (Singh & Khilar, 2017). They estimate the location of the nodes through network connectivity and with the help of anchor nodes (nodes having their own location information). The range free localization algorithms provide cost effective localization methods, but the results are less precision and accuracy. However, constrained WSNs adopt range free localization methods as the most preferred localization solution and research is in progress to further improve localization accuracy achieved by these algorithms. Some typical range free algorithms are Approximate Point in Triangle (APIT) (He, Huang, Blum, Stankovic, & Abdelzaher, 2003), Mid Perpendicular (Singh & Khilar, 2017), Distance Vector-Hop (DV-Hop) (Niculescu & Nath, 2001).
In range free algorithms, DV-Hop is the most popular method. DV-Hop is a distributed, hop-by-hop positioning algorithm. The algorithm is a three steps procedure viz. (1) acquiring hop count between anchor nodes, (2) calculating average hop size and finally, location of sensor nodes are estimated by trilateration method (Niculescu & Nath, 2001; Asmaa, Hatim, & Abdelaaziz, 2014). DV-Hop is one of the simplest, cost effective and scalable algorithms with high coverage quality. But localization error in DV-Hop algorithm is very high due to presence of ranging errors. In literature, several modifications to DV-Hop algorithm are proposed and research is still in progress to reduce computational complexity and increase the localization accuracy.