Positioning systems using wideband radio signals are able to maintain centimeter-level accuracy if clear line-of-sight measurements to several base stations are available. Obstructed or blocked paths in addition to harsh multipath propagation limit the achievable accuracy. As a remedy, numerous base stations are employed which are associated with increased deployment costs.

In this thesis we tackle localization using radio signals with a minimum amount of available infrastructure, e.g. limited or no access to base stations. To remedy a potential performance loss, three research directions are identified. First, we formulate a site-specific double-directional channel model to enhance the potential of exploiting multipath propagation for localization. We examine propagation mechanisms like path reflections or diffuse multipath with respect to their contained position information. Available prior knowledge of the surrounding environment enables to predict the distributions of both specular reflections as well as diffuse multipath in angle-delay domain. This enables to perform localization using a single base station even in environments with a considerable large amount of path blocking.

The second research direction treats exploitation of measurements with respect to angle and delay information using directive antennas and non-phase-coherent transceivers. Directive antennas attenuate radio signals as function of their transmitting/receiving angles. Having a number of directive antennas, the antennas’ directivity enables to exploit multipath angle information without the need of cost-intensive, coherent multi-antenna transceivers. In comparison to an antenna, radiating in an isotropic manner, our theoretic analysis demonstrates the advantage of directive antennas to separate multipath in angle-delay domain, by reducing multipath interference.

The third research direction points to cooperative localization. We demonstrate that a site-specific multipath channel model provides global position coordinates. The users’ position coordinates are estimated from specular reflections originating at reflective surfaces. Prior knowledge of the surface location enables to determine the users’ positions without requiring a base station at all.