• Cramér-Rao Lower Bound
• CRLB
• crowdsourcing
• indoor localization
• IP geolocation
• localization accuracy

HIS Ph.D. thesis focuses on network-based localization techniques in two different scenarios: localization of host belonging to the Internet (IP geolocation), and indoor localization of host moving within buildings. Independently from the specific scenario, we advocate the use of constraints (e.g. distance-based, time-based) for improving the localization accuracy. The proposed IP geolocation technique, differently from existing literature, uses smartphones as landmarks to perform measurements from which distances are estimated. Measurements collected in wireless scenarios are noisier than the ones collected in wired scenarios. The use of constraints (time-based, distance-based, access technology-based) reduces the effects of noise, thus improving the accuracy of localization. For the indoor scenario, this thesis proposes a method based on Wi-Fi and external temporal constraints (schedules) to reduce the ambiguity of localization. The approach has been tested in a real scenario and results show that observing users’ movements is sufficient to infer their schedules. The last contribution of this thesis is to derive the Cramér-Rao Lower Bound (CRLB) for IP geolocation.
The CRLB has been widely used to define the maximum theoretical accuracy of localization methods. However, as far as we know, it is here used for the first time in the contest of IP geolocation. If the performances of a localization system are close to the CRLB, this means significant improvements are difficult to achieve.