Abstract

Wireless localization (WL) has gained more popularity in the recent years due to large developments in the field of autonomous driving, as well as unmanned aerial vehicles (UAVs) and robotics. Especially in the field of autonomous driving where vehicles can be modeled by rigid bodies, with a known shape, instead of a single point, rigid body local- ization (RBL) offers great advantages in terms of yielding not only the targets position, but also its orientation. In that sense, this work gives an insight into simple state-of-the-art (SotA) single-point localization methods, leading to rigid body localization (RBL) techniques, estimating the translation and rotation of a rigid body moved in space. The proposed method of this thesis offers an extension of the multidimensional scaling (MDS)-based SotA in an an- chorless scenario, estimating not only the relative translation between two rigid bodies, but the complete translation vector. To counter the effect of incomplete observations in the ranging measurements due to non-line-of-sight (NLOS) paths, an intensive discussion about the history of matrix com- pletion algorithms is presented with a discussion about modern, so called, discrete-aware methods, and their usage in RBL frameworks. Finally, simulation results illustrate the superior performance of the proposed methods against the SotA in terms of the root mean square error (RMSE), followed by a conclusion of the work and an outlook into future research directions.