Previous Abstract Return to Session C3 Next Abstract

Session C3: Positioning with Non-GNSS Radio Signals (Invited Session)

Enhanced Multilateration Methods with a Global Approach
Rabih Chrabieh, Mazen Neifer, Ganda Ouedraogo, Ines Ben Hamida, Peter Bagnall, Nestwave, France
Location: Atrium Ballroom
Alternate Number 1

Modern wireless cellular systems such as LTE (4G) provide various mobile positioning methods that use received power or time of flight (distance) estimation between a User Equipment (UE) and a set of cellular antenna towers in order to locate the UE. In particular, a standardized downlink TDOA (Time Difference of Arrival) method in LTE is denoted by OTDOA (Observed Time Differences of Arrival) and uses the time of arrival (TOA) of reference signals (RS) or pilots in order to determine time of flight between the UE and each antenna tower. The user equipment (UE) estimates the RS time of arrival from neighbor cells relative to the RS time of arrival from a reference or serving cell. The difference between the two measured time of arrivals is denoted by RSTD (Reference Signal Time Difference).
The UE reports several estimated RSTD measures between neighbor cells and a reference cell. The measurements are reported to a location server that knows the position of the antenna towers. The combination of antenna tower positions and RSTD measurements enable locating the UE via some multilateration formula. The multilateration problem typically consists of minimizing a Least Squares problem, the solution of which are the UE coordinates.
The problem with the TDOA or RSTD estimation method is that it ignores an important parameter or observation that can improve the solution, especially when the TOA measurements of the reference cell are noisier than those of other cells (e.g. outlier measurement).
On the other hand, in satellite navigation systems (GNSS), the multilateration problem is stated as a Time of Transmission problem or TOT. In this case, the time of arrival is theoretically defined in the system model as theoretical time of flight plus UE clock bias plus receiver noise.
The problem with systems using a TOT method is that current art is not in the form of a closed-form, global solution, thereby limiting convergence speed and robustness in satellite navigation systems and terrestrial-based systems.
In this paper, we show how current solutions such as those used in mobile wireless networks can be improved by transforming the TDOA problem into a TOT problem, and solving the problem using a global solution, taking into consideration UE’s clock bias, as well as the UE’s Doppler.
The global solutions are closed-form solutions with approximate weights. Unlike the local Taylor expansion solution, they require no knowledge of an approximate initial guess. In TDOA multilateration of terrestrial systems (e.g. OTDOA, WiFi TDOA, LoRa TDOA), a good initial approximate guess is typically unknown. In GNSS systems, an initial guess can be chosen as the center of the Earth; but in some cases, it can lead to an invalid local solution (i.e. local minimum) when the first order approximation is not good enough.
We then provide improvement to the LTE OTDOA solution by defining new information that can be used by the UE or transmitted by the UE to the location server in order to improve multilateration accuracy.
We present novel formulations and global solutions to the TOT multilateration problem using a UE clock bias that are superior to the well-known TDOA’s reference node multilateration methods; i.e. we improve upon Spherical Interpolation, Spherical Intersection, and Constrained Spherical Interpolation. We further extend the solution to the Least Absolute Deviations criterion in order to better handle measurements in Non-Line-of-Sight or in heavy multipath scenarios.
We further propose extension to LTE OTDOA such that the UE can report Doppler estimates, in particular when measurements are spaced out in time.
Simulation and field results show the improvement obtained by replacing a TDOA problem with a TOT problem, solving the TOT problem using the novel global solution, and substituting the LS criterion by a LAD criterion in difficult environments such as urban canyons or indoor.



Previous Abstract Return to Session C3 Next Abstract