RSS based localization in a rayleigh fading environment

The objective of this thesis is to quantify the improvement that can be obtained in sensor agent localization accuracy as a function of the number of multipath components that can be resolved. We assume that a known number sensor agents are located at unknown coordinates within a rectangular grid having anchors at the corner locations, whose locations are known. Further, we assume fading is Rayleigh and that the propagation constant is constant but unknown. Also, we assume that modulation is spread spectrum and that either the sensors or agents are capable of resolving multipath components down to the chip level and are capable of measuring the received signal strength in each of the resolved multipath components. An error function is formulated based upon the square of the distances between the actual sensor locations and their model-predicted locations, which are functions of the received signal strength of the various multipath components and the propagation constant, and the optimal sensor location estimates and propagation constant are determined through a multistage process of formulating and minimizing error functions. The effectiveness of this approach is investigated via extensive simulations in which the Saleh-Valenzuela model is used to generate multipath components. The simulation results indicate that for a given fixed propagation constant, resolving multipath results in improved localization accuracy and that this improvement is a non decreasing function of the propagation constant. For a distance-squared propagation environment, the results indicate that resolving 6 multipath components improves localization accuracy by at least 20 %, the improvement being with respect to the localization accuracy based on aggregate received signal strength.