Localization strategies for large-scale airborne deployed wireless sensors

Abstract

Localization is the process of finding the geometric locations of wireless sensor nodes according to some real or virtual coordinate system. It is an important task when direct measurements of the wireless sensor locations are not available. From the various techniques evolved in localizing sensor nodes, one approach is to use the received signal strength to predict the location of unknown sensing devices. In this paper, passive localization algorithms are developed, presented and tested. The algorithms perform region-based localization of stationary wireless sensors with respect to a frame of reference using received signal strength of the sensors. The reported work is conducted in two phases, theoretical development then simulation and hardware testing. In the first phase, localization algorithms were developed to predict the location of wireless sensor nodes. We categorized localization of sensors in three different classes. In class-I, localization is done for sensors that are in the communication range of at least three head nodes. In class -II, localization is done for sensors in the communication range of two head nodes, and in class-III, localization is done for sensors that are in the communication range of only one head node. In the second phase, the three different categories were tested by simulation then using hardware. A test-bed was established using the crossbow (MICAz) hardware and used to measure the sensors transmission signal strength. Then, the localization software provided with these signal strength as input to predict the location of each wireless sensor nodes. The algorithm developments, the simulation and hardware preliminary test results of the localization algorithms are presented in this paper.