Improving GPS Code Phase Positioning Accuracy in Urban Environments Using Machine Learning

Abstract

The accuracy of location information, mainly provided by the global positioning system (GPS) sensor, is critical for Internet-of-Things applications in smart cities. However, built environments attenuate GPS signals by reflecting or blocking them resulting in some cases multipath and non-line-ofsight (NLOS) reception. These effects cause range errors that degrade GPS positioning accuracy. Enhancements in the design of antennae and receivers deliver a level of reduction of multipath. However, NLOS signal reception and residual effects of multipath are still to be mitigated sufficiently for improvements in range errors and positioning accuracy. Recent machine learning-based methods have shown promise in improving pseudorange-based position solutions by considering multiple variables from raw GPS measurements. However, positioning accuracy is limited by low accuracy signal reception classification. Unlike the existing methods, which use machine learning to directly predict the signal reception classification, we use a gradient boosting decision tree (GBDT)-based method to predict the pseudorange errors by considering the signal strength, satellite elevation angle and pseudorange residuals. With the predicted pseudorange errors, two variations of the algorithm are proposed to improve positioning accuracy. The first corrects pseudorange errors and the other either corrects or excludes the signals determined to contain the effects of multipath and NLOS signals. The results for a challenging urban environment characterized by high-rise buildings on one side, show that the 3-D positioning accuracy of the pseudorange error correction-based positioning measured in terms of the root mean square error is 23.3 m, an improvement of more than 70% over the conventional methods.

Machine Learning-Based Short-Term GPS TEC Forecasting During High Solar Activity and Magnetic Storm Periods

“…The gradient boosting decision tree (GBDT) approach combines regression trees with a gradient boosting technique, which has been widely used to solve regression and classification problems such as soil moisture estimation [33], GNSS positioning accuracy improvement [34] and GPS signal reception classification [35]. It achieves the best performance when comparing with the other classic machine learning techniques [35], [36].…”

Application of Machine Learning in Electromagnetics: Mini-Review

“…ML techniques have been significantly applied in estimating antenna position and direction to achieve the maximum gain in transmission and receiver systems. It assists in the detection and control of beam phasing of antennas based on signal patterns, strength, and target location [79,80]. The direction of arrival (DoA) estimation has become a popular application in military-civil research for remote object detection [81][82][83].…”

Semantic VPS for Smartphone Localization in Challenging Urban Environments

“…We used the Sørensen-Dice coefficient metric to compare the region of two material segmented images [29]. Equation (8) shows the calculation of the similarity index for each material class:…”

“…In the context of outdoor pedestrian localization, the application of the global navigation satellite system (GNSS) is key to providing accurate positioning and timing services in open field environments. Unfortunately, significant improvement is needed in the positioning performance of GNSS in urban areas due to signal blockages and reflections caused by tall buildings and dense foliage [ 8 ]. In these environments, most signals are non-line-of-sight (NLOS), which can severely degrade the localization accuracy [ 9 ].…”