In wireless sensor networks, localization is one of the essential requirements. Most applications are of no use, if location information is not available. Based on cost, localization algorithms can be divided into two categories, namely range-based and range-free. Range-free are cost-effective, but they lack accuracy. In this chapter, the role of nature-inspired algorithms in enhancing the accuracy of range-free algorithms has been investigated. Inferences drawn from exhaustive literature survey of recent research in this area establishes the importance of these algorithms in sensor localization.
TopIntroduction
Wireless Sensor Networks (WSNs) is hot research area nowadays (Rawat, Singh, Chaouchi, & J.M.Bonnin, 2014). It is collection of sensitive and dedicated sensor nodes that sense some critical event such as heat, moisture, light, pressure, etc and processes this information and sends it to base station via multiple hops (Kaur, Kumar & Gupta, 2019). WSN can be used for many critical applications such as military, forest surveillance, and intrusion detection, civilian applications because of its rapid deployment, self organizing behavior and fault tolerant nature. In all these kinds of applications, main role of WSN is to detect an event and send it to base station for immediate action (Han, Xu, Duong, Jiang & Hara, 2013). But for immediate response, location of event is also required. This identification of location of event i.e. location of sensor that detected that vent is called Localization. This is most challenging problems in field of WSN. The technique makes use of few sensor nodes that are usually deployed uniformly in whole WSN and have knowledge of their locality. With the aid of these anchor nodes, remaining unknown nodes try to find their location. Localization contains two phases: Distance estimation and Position estimation as depicted in Figure 1 (Farooq-i-Azam & Ayyaz, 2016).
In localization process, firstly unknown nodes predict their distance from anchor nodes and then predict their location. Distance/Angle estimation process can be of two types depending upon measurement technique and are range-based or range-free (Zhao, Xi, He, Liu, Li & Yang, 2013). Range based attains distance/angle data among neighbour nodes and usually uses specialized equipment, whereas Range free applies connectivity data between neighbouring nodes without using any specialized equipment. Examples of range based include Received Signal Strength Indicator (RSSI) (Girod, Bychobvskiy, Elson & Estrin, 2002), Time of Arrival (TOA) (Harter, Hopper, Steggles, Ward & Webster, 2002), Time Difference of Arrival (TDOA) (Cheng, Thaeler, Xue & Chen, 2004) [8], Angle of Arrival (AOA) (Niculescu & Nath, 2003), etc. Examples of Range free are Centroid (Bulusu, Heidemann & Estrin, 2000), Distance Vector-Hop (DV-Hop) (Niculescu & Nath, 2001), Amorphous (Nagpal, Shrobe & Bachrach, 1999), Approximate Point in Triangle (APIT) (He, Huang, Blum, Stankovic & Abdelzaher, 2003), Multi Dimensional Scaling (MDS) (Ruml & Shang, 2004), etc. Range free is more popular due to its cost effectiveness as they do not require any expensive hardware unlike range based techniques. But range free has low accuracy. There have been so many proposals to improve its accuracy in last two decades. This paper surveys all these proposals that uses nature inspired algorithms for improvement.