Improvement in Sensor Node Localization

Abstract

The most fundamental problem of wireless sensor networks is localization (finding the geographical location of the sensors). Most of the localization algorithms proposed for sensor networks are based on Sequential Monte Carlo (SMC) method. To achieve high accuracy in localization it requires high seed node density and it also suffers from low sampling efficiency. There are some papers which solve this problems but they are not energy efficient. Another approach the Bounding Box method was used to reduce the scope of searching the candidate samples and thus reduces the time for finding the set of valid samples. In this thesis we propose an energy efficient approach which will further reduce the scope of searching the candidate samples, so now we can remove the invalid samples from the sample space and we can introduce more valid samples to improve the localization accuracy. We will consider the direction of movement of the valid samples, so that we can predict the next position of the samples more accurately, hence we can achieve high localization accuracy. Further we can also add the information about the speed of movement of the node so that we can measure the actual acceleration of the node. Now as we have information about the direction and speed of movement of the node we can locate a sensor node more accurately and faster.