[Back]

D. Pradipta, G. Retscher:
"Feasibility Study of the Possibilities to Reduce the Multipath Bias for Smartphone Positioning in Urban Environments";
Poster: IAG Scientific Assembly 2013, Potsdam, Germany (invited); 2013-09-01 - 2013-09-06; in: "IAG Scientific Assembly", (2013), 2 pages.

Personal navigation systems intend to provide the navigation information in any environment, indoors and outdoors, and at any time. Thereby in outdoor environment positioning is usually provided by using GPS. One of the challenging application areas is smartphone positioning which is very popular in Location-based Services (LBS) and other navigation tasks. However, GPS is inaccurate or even fails in urban/dense areas because of several problems such as loss of satellite signals and multipath effect. Therefore several methods have been introduced to improve the navigation accuracy.
One of the main problems in GPS positioning is multipath which additionally limits the performance of personal navigation systems especially in challenging environments such as in forests or dense areas in urban canyons. Multipath occurs when signals from a satellite arrive at the receiver multiple times, this is usually caused by reflection of the signal from nearby buildings or objects depending on which frequency is used. There are several options to reduce the multipath, i.e., choosing locations that are not likely have reflective surfaces nearby, using special antennas, and/or setting proper elevation cut off angles. Such strategies, however, cannot be followed in every situation and circumstance. Therefore as new positioning applications have emerged increasing demand for eliminating the effect of multipath arose. In addition, due to the complexity and cost of the measuring hardware the use of other external information sources to improve the position is needed.
A possible solution of this problem is to investigate the multipath effect from related sources and introduce it to improve the positioning accuracy. Depending on the application, multipath can be considered as a systematic error when in repeated measurements under the same conditions the error occurs. In either case, linear combinations that reduce the noise often reduce the multipath as well. For the experiment, we will investigate if PPP (Precise Point Positioning) measurements using a geodetic GPS system can be used for multipath effect estimation. The estimated multipath bias shall then be applied to measurements of different smartphones for positioning along a reference trajectory in urban environment. The trajectory in the surroundings of the Vienna University of Technology is around 1.1 km long consisting of 24 reference points which have been determined using classical surveying methods with a totalstation. Therefore the local coordinates of these reference points can be considered as ground truth. From PPP with the geodetic GPS receiver position corrections are derived and applied to the positions obtained from the smartphones to reduce the multipath bias. In addition, simulations are performed to investigate the feasibilty of this approach. The spectrum of different conditions such as satellite visibility and geometry during a day is also considered. In addition, repeated measurements during same conditions on several days are performed. The aim of these experiments is to analyze how frequent an update and how long observation times are needed to be able to model the multipath and to eliminate or reduce the multipath effect in the measurements. For this reason, an analysis into methods that reduce multipath is conducted in this study.