Multiple Sources Localization Using TDOA And GROA

Passive source localization is a fundamental problem that has found numerous applications in electromagnetic spectrum monitoring, wireless communications, sensor networks, microphone arrays and many others. The localization technology based sensor networks has not only provided a positional information service, urgent rescue and goods tracing function for the customers, it has also played a very effective role in national security and crime prevention. This thesis performs a fundamental investigation of whether the gain ratios of arrival (GROAs) can be utilized in conjunction with the time differences of arrival (TDOAs) to improve the multiple sources localization accuracy with erroneous sensor positions. The new algorithm will provide the more accurate and more reliable location result for the applications mentioned above.Firstly, this thesis will introduce the application of wireless localization technology in sensor network. It will discuss the sensor network system and different kinds of regular usages in localization technology. And then, this thesis considers the problem of locating multiple disjoint sources simultaneously, and it performs a fundamental investigation of whether the gain ratios of arrival (GROAs), defined here as the ratio of the received signal amplitudes at the referenced sensor to the other sensors, can be utilized in conjunction with the time differences of arrival (TDOAs) to improve the source localization accuracy with erroneous sensor positions. This is a challenging problem and closed-form solution with good localization accuracy need to be found. This thesis proposed an algorithm that jointly estimates the unknown sources and sensor positions to take the advantage that the TDOAs and GROAs from different sources have the same sensor position displacements. We use the idea of hypothesized source locations in the algorithm development to enable the formulation of psuedolinear equations, thereby leading to the establishment of an closed-form solution for multiple sources location estimates using GROAs and TDOAs. For clarity, the localization of two disjoint sources is used in the algorithm development. We also derive the CRLB of multiple source location estimate using both RDOAs and GROAs when sensor positions have errors.Numerical simulations are included to support and corroborate the theoretical developments. Our conclusion is that MSE of the proposed algorithm has at least10dB improvement compared with algorithm using TDOAs only when the signal bandwidth is large than8.5KHz and SNR is large than-4dB. Most importantly, the new method is shown analytically to achieve the CRLB accuracy for Gaussian data model at moderate noise level.