High Accuracy Terrestrial Positioning Based on Time Delay and Carrier Phase Using Wideband Radio Signals

Abstract

Accurate position solutions are in high demand for many emerging applications. Global navigation satellite systems (GNSS), however, may not meet the required positioning performance, especially in urban environments, due to multipath and weak received power of the GNSS signal that can be easily blocked by surrounding objects. To achieve a high ranging precision and improve resolvability of unwanted reflections in urban areas, a large signal bandwidth is required. In this thesis, a terrestrial positioning system using a wideband radio signal is developed as a complement to the existing GNSS, which can provide a better ranging accuracy and higher received signal power, compared to GNSS. In the terrestrial positioning system presented in this thesis, a wideband ranging signal is implemented by means of a multiband orthogonal frequency division multiplexing (OFDM) signal. All transmitters are synchronized by time and frequency reference signals, which are optically distributed through the white-rabbit precision time protocol (WR-PTP). Like in GNSS, the to-be-positioned receiver is not synchronized to the transmitters. Positioning takes place through range measurements between a number of transmitters and the receiver. Time delay and carrier phase are to be estimated from the received radio signal, which propagated through a multipath channel. This estimation is done on the basis of the channel frequency response and using the maximum likelihood principle. To determine whether or not reflections need to be considered in the estimation model, a measure of dependence is introduced to evaluate the change of the precision (i.e., variance), and the measure of bias is introduced to assess the bias of the estimator when the reflection is not considered. Also, a methodology is proposed for sparsity-promoting ranging signal design in this thesis. Based on a multiband OFDM signal, ranging signal design comes to sparsely select as few signal bands as possible. Using fewer signal bands for ranging leads to less computational complexity in time delay and carrier phase estimation, while the ranging performance can still benefit from a large virtual signal bandwidth, which is defined by the entire bandwidth between the two signal bands at the spectral edges. It is proposed to use the Cramer-Rao lower bound (CRLB) of time delay estimation, the measure of dependence, and the measure of bias as constraints in ranging performance, and formulate an optimization problem to design a sparse multiband signal.