The effect of pruning stage in secure localization in Wireless Sensor Networks

In Wireless Sensor Networks (WSNs), the received data from nodes are useful only when accurate information about their locations is available. Many localization methods use some nodes that their locations are known by manual or by means of positioning systems such as GPS. These nodes are known as anchor or beacon nodes. In those applications of WSNs in which localization is very important, such as robotic land mine detection, battle field surveillance, etc., the anchor nodes may be attacked by unauthorized accesses in hostile environments and as a result, the localization system may face perturbation. Therefore, in recent literatures some secure localization methods have been introduced. Gradient descent is a low complex, secure and time efficient approach that marks some anchor nodes as attacked ones by using special algorithms and prunes them, then determines the position of sensors using the remaining anchor nodes. In this paper, it is shown that pruning these anchor nodes may cause to localization error increase. In other words, it is shown that in non-coordinated attacks, using the information sent by attacked anchor nodes can improve the localization accuracy. The simulation results show that by utilizing this technique, we can decrease the localization error up to 15%.