Authors:
Neunteufel, Daniel
Abstract:
Indoor localization has become an important topic in recent years. While for outdoor use, highly developed global navigation satellite systems (GNSS) are the omnipresent method for localization, their application in indoor scenarios is limited, due to obstructed line-of-sight (LoS) links to the satellites. Despite many years of vigorous research, no comparable single dominant method has been established for indoor use cases. Rather, there exist a multitude of different approaches, from both hardware and algorithmic perspective, all having indiv (...)
Indoor localization has become an important topic in recent years. While for outdoor use, highly developed global navigation satellite systems (GNSS) are the omnipresent method for localization, their application in indoor scenarios is limited, due to obstructed line-of-sight (LoS) links to the satellites. Despite many years of vigorous research, no comparable single dominant method has been established for indoor use cases. Rather, there exist a multitude of different approaches, from both hardware and algorithmic perspective, all having individual advantages and disadvantages. Suitable solutions often strongly depend on the concrete application. For example, tracking of a single moving user is different from location awareness of large ensembles of internet of things (IoT) nodes. The latter is of particular research interest, as over the years vast numbers of such nodes were deployed for numerous applications, initially without capabilities for localization. Over time, new use cases might arise for such nodes which have not been foreseen upon deployment. The reuse of existing IoT hardware is desirable in such cases, as changes to large numbers of nodes are costly. Possibilities to enable localization under these circumstances are studied in this thesis. The focus lies on the use of radio waves in the industrial, scientific, and medical (ISM) radio band at 2.4 GHz, as many existing systems make use of this unlicensed frequency band. Multipath fading and interference is characteristic for the typical indoor channel. To overcome the drawbacks of the limited signal bandwidth of widely used hardware in scenarios prone to such multipath effects, a novel method is suggested which allows extending the available transmit signal bandwidth. With this method, meaningful time of flight (ToF) measurements become feasible. It relies on direct manipulation of the synthesizer of a transceiver chip, allowing for chirped signal modulation. This signal modulation and the required processing arising from the proposed non-standardized approach is elaborated. Unlike the transmission bandwidth, the receiver bandwidth of the transceiver chips cannot be increased similarly. Thus, a dedicated receiver infrastructure with a sufficient sampling rate is required to facilitate the localization of transmit-only nodes. As synchronization of nodes and infrastructure is assumed impossible, time difference of arrival (TDoA) calculations for synchronized receiving anchors at known locations are studied. For such an implementation with synchronized receivers, a suitable system model is introduced, taking into account a possibly non-ideal synchronization. Using the system model and thereof derived estimation bounds, the requirements on such an infrastructure are assessed. During a measurement campaign, the proposed method has been tested under real-world conditions, using a software-defined radio (SDR) platform. A combination of TDoA with angle of arrival (AoA) and received signal strength (RSS) measurements allows achieving a root-mean-square error (RMSE) of 2.19. This proves that good localization results are achievable even with limited hardware capabilities.
(Read More)