Search

Global Navigation Satellite Systems (GNSS) are systems that continuously acquire data and provide position time series. Many monitoring applications are based on GNSS data and their efficiency depends on the capability in the time series analysis to characterize the signal content and/or to predict incoming coordinates. In this work we propose a suitable Network Architecture, based on Long Short Term Memory Recurrent Neural Networks, to solve two main tasks in GNSS time series analysis: denoising and prediction. We carry out an analysis on a synthetic time series, then we inspect two real different case studies and evaluate the results. We develop a non-deep network that removes almost the 50% of scattering from real GNSS time series and achieves a coordinate prediction with 1.1 millimeters of Mean Squared Error.

La rete di stazioni GNSS permanenti attualmente presente in Italia è costituita da più di 700 siti. Queste stazioni sono distribuite in modo non omogeneo sul territorio e sono caratterizzate da una forte disomogenietà; infatti possono differire tra loro per diversi motivi, quali ad esempio: le caratteristiche della strumentazione utilizzata; la collocazione dell'antenna ricevente (roccia, edificio); il tipo di supporto antenna utilizzato (tripode INGV, supporto UNIBO-Max, staffa), gli scopi per cui è stata realizzata la stazione (studio e monitoraggio scientifico, supporto per il posizionamento in tempo reale o altro). Questi fattori possono influire pesantemente sulle caratteristiche delle serie temporali giornaliere della posizione introducendo segnali periodici e non ed influire in modo non trascurabile sulle caratteristiche del rumore presente nelle serie temporali della posizione di queste. Per cercare di individuare possibili correlazioni tra questi fattori e le caratteristiche delle stazioni permanenti italiane, le serie temporali della posizione ottenute elaborando le osservazioni giornaliere con due diversi software (Gipsy/OASIS e Gamit /Globk) sono state analizzate. In particolare, è stata utilizzata l'analisi spettale per individuare i principali segnali periodici presenti nelle serie temporali della posizione. Inoltre, le serie temporali sono state analizzate per determinare le caratteristiche del rumore presente per tentare di correlarlo con le diverse tipologie di installazione e/o gestione presenti nella rete rete italiana. E' stata anche utilizzata un'analisi di tipo principal component analysis (PCA) allo scopo di individuare possibili segnali comuni a stazioni presenti in determinate aree o a siti che presentano tipologie di installazione o gestione simili.

Al giorno d’oggi, la tecnologia GNSS può essere uno strumento utile, non solo in ambito di navigazione, ma anche per applicazioni di posizionamento di precisione come il monitoraggio di strutture o del territorio. Questa tecnica permette un controllo in continuo, anche da remoto, e ha un costo relativamente basso sia per quanto riguarda le strumentazioni che per gli aspetti di materializzazione e mantenimento. Nell’ottica di utilizzare la tecnologia GNSS per applicazioni di monitoraggio, il problema, nella sua interezza, può essere suddiviso in due grandi ambiti: l’ambito tecnologico (strumentazioni e apparati da adottare per un problema di monitoraggio anche in tempo reale) e l’ambito scientifico legato al problema del trattamento del dato che può, o deve, essere separato a sua volta in due aree legate alla tipologia di monitoraggio che si intende compiere. Il monitoraggio di movimenti lenti per analisi di stabilità (volto principalmente allo studio dell’evoluzione del quadro deformativo nel lungo periodo) o il monitoraggio per possibili applicazioni nell’ambito dell’early warning (che richiedono la valutazione di un movimento in tempo reale). Per le sperimentazioni sia di natura tecnologica che di trattamento del dato sono stati individuati due siti su cui sono stati installate stazioni permanenti GNSS di classe geodetica. Il primo sito è collocato sul tetto della scuola di Ingegneria ed Architettura dell’Università di Bologna, il secondo sito è la sommità della Torre Garisenda di Bologna (oggetto di particolare rilevanza culturale). Sul tetto di Ingegneria è stata collocata anche una slitta micrometrica che ha consentito l’imposizione di movimenti calibrati nelle tre direzioni. Per ciascun aspetto sono stati compiuti numerosi test sperimentali volti ad individuare le tecniche migliori per il monitoraggio con GNSS a partire dalle problematiche di trasferimento dati e finendo con lo sviluppo di particolari algoritmi per migliorare le prestazioni del sistema in condizioni di monitoraggio statico., Nowadays, GNSS technology can be a useful tool, not only for navigation and location aspects, but also for precise positioning applications, such as structures and environmental monitoring. This technique allows a continuous control, even remotely, and has a relatively low cost both regarding the instrumentation, the monumentation and the maintenance. In order to use the GNSS technology for monitoring applications, two main problems can be identified: (a) the technological one is represented by instrumentation and equipment that are used in order to provide data in real time. The scientific one (b) is related to the issue of data processing. It can be in turn separated in two fields, which are related to the type of monitoring. In detail, (b1) the monitoring of slow movements for stability analysis (aimed mainly at the study of the long time evolution of these phenomena) or (b2) the monitoring for possible applications in early warning systems. For this study two sites were considered. In each one, a GNSS geodetic class permanent station was installed. The first site is located on the roof of the School of Engineering and Architecture, University of Bologna, while the second one is on the top of the Garisenda Tower in Bologna. Moreover, in the first site, a micrometric sledge which allowed the imposition of calibrated movements in the three directions, has also been placed. For each aspect, several experimental tests have been carried out. They were aimed at identifying the best techniques for GNSS monitoring, starting from the data transfer problems and ending with the development of algorithms to improve system performance in static monitoring conditions.

The present Italian kinematic field has been estimated by a network of about 780 GNSS permanent stations. The daily observations acquired with a sampling rate of 30 seconds from January 01, 2001 to December, 31, 2015 have been analysed through the GAMIT/GLOBK software, adopting a distributed procedure. The position component time series (North, East and Vertical) are analysed in order to obtain a reliable estimation of: average velocity, amplitude and periods of seasonal signals, dimension of instrumental and/or seismic steps and noise. These characteristics have been studied in order to verify possible correlations with some characteristics of GNSS installation and/or natural and anthropogenic phenomena present in some areas. The average velocity of the 414 sites characterized by a time span greater than 2.5 years and an efficiency greater than 50% have been used in order to estimate the present kinematic pattern in the Italian peninsula and surrounding areas. The horizontal velocity field is characterized by some interesting features that are compatible with the long term deformation pattern deduced from geological observations. The vertical velocity field is characterized to an heterogeneous distribution probably due to the strong influence that anthropogenic activities have on this component .

Nowadays, GNSS technology can be a useful tool, not only for navigation and location aspects, but also for precise positioning applications, such as monitoring of structures. This technique permits an all-weather continuous remote control as well as a quite low cost, both for instrumentations and for monumentation aspects. The traditional approach to GNSS monitoring is based on a relative positioning between a reference station, assumed as stable, and a rover one. Rover and reference stations can be situated at short distance, in order to obtain high accuracy with L1 receivers. For data processing, several software packages are available, both by GNSS manufacturers and by scientific institutions. Recent free and open source software packages, such as RTKLIB and GoGPS, are particularly interesting because they allow a much higher level of customization of the data processing and post processing phases and, for short distances, they offer performances comparable with those of more consolidated softwares, such as Bernese, Gamit, etc… The continuous monitoring based on a daily position estimation produces a time series that can be useful both for structure stability studies and for “early warning” systems. In this paper, several aspects related to the monitoring problem using GNSS have been investigated using a GNSS receiver located on the top of the Garisenda tower (Bologna, Italy) since October 2013. oriously affected by problems of stability and has already been monitored using different techniques, such as inclinometer, spirit levelling etc... The acquired data, amounting to a couple of years, have been processed starting from four GNSS reference stations located in different parts of Bologna, but each one at a maximum distance of 2km from the tower. Assuming each reference station as stable, a single base data processing has been performed and four time series have been obtained. After a phase of outliers detection and removal, mean velocities for each time series have been estimated using a least squares approach. A comparison of the mean velocities shows significant differences, probably due to some instabilities of the reference stations. Therefore, each reference station has been processed and a model for each reference station has been defined using Precise Point Positioning approach; the solutions have been aligned to the ETRF using local parameters transformations, generated by a regional network of 16 stations that covers the Italian peninsula and the bordering Countries. The models obtained have been considered for a new data processing and new mean velocities of the Garisenda station have been estimated. The new time series have a higher level of accuracy than the previous series. Despite that improvement in the level of agreement between the four solutions, some differences are still present. This work reports a proposal concerning more complex analysis, with the aim to improve solutions and results interpretation.

In this paper the main results obtained by a study on a subnetwork of the EUREF Permanent Network has been reported. The study focus the impact of different methods to align PPP solution into ITRS Reference System using both global and regional transformation parameters. In particular we demonstrate as the regional parameter transformation improve accuracy and repeatability without remove any local periodical signal detected by the permanent stations. The considered dataset is constituted by 6 years of 14 permanent stations located in Italy and neighbor regions. This work is a review of a more complete work published in 2015 in GPS solution entitled “Improved PPP performance in regional networks” by the same coauthors of this contribute.

GNSS technology is still not considered as a suitable method for structural monitoring because of its relatively low precision, despite the increase in acquisition frequencies. This paper aims to evaluate a strategy for filtering a daily kinematic GNSS solution using a movement smoothing model based on the observations of previous days, so as to obtain a less scattered solution. A test was conducted with a permanent GNSS station located on top of the medieval Garisenda tower in Bologna (Italy), as it is an important part of the city’s cultural heritage and a reliable test site. Because of the presence of the adjacent Asinelli tower, which is taller than the Garisenda, sky visibility is not optimal here, and for this reason a particular sequential filtering can be adopted in order to obtain more accurate solutions. The test was performed using RTKLIB software to calculate 1 Hz baselines between the test station and a master one located on a stable area about one kilometer away from the tower. In order to obtain reliable results, several variables were considered both in the data processing phase and in defining the filter. All results are reported and discussed in detail in the paper. The test results reveal a reduction in scatter of about 20% in the filtered kinematic time series, especially in weaker geodetic components.

The increasing number of GNSS permanent stations both in Italy and around the world can have a huge amount of data that may be monitored continuously and in accurate and precise wise. From some years the calculation approach known as Precise Positioning Point (PPP) constitutes a valid alternative to the differential approach, since it allows to obtain precision and accuracies of the same order of magnitude but in the face of a flexibility of calculation certainly advantageous and computation time significantly reduced. These accuracies are reachable applying an ambiguity resolution strategy and a successive framing approach. As known PPP estimate coordinates of points (GNSS stations) into the same reference frame (RF) of the orbits that can be different from the International or National RF (IGS08 or others). In this work, the computing architecture that allows to manage the monitoring of several networks of permanent stations is shown. This procedure is already running and it will be shown from a 16 GNSS permanent stations network evenly distributed on Italy and neighboring areas in a time window of 11 years.

GNSS is one of the most used technique to monitor Reference Frame, Plate motion, Landslides, Structures, etc... Data acquired by GNSS satellites have to be processed considering models, approaches and strategies which are influencing the final results and, if wrongly done, may introduce biases. In this work the Precise Point Positioning (PPP) approach for GNSS data processing has been considered and investigated. As well-known PPP is faster and more flexible compared to the carrier phase differenced approach. On the other hand the differenced approach has been generally characterized by a better repeatability, especially for quite small baselines. Until few years ago one of the major limits of PPP was the impossibility to resolve the ambiguity as integer but nowadays some methods are available to resolve this aspect expecting an increasing of the precision. Starting from two subsets of the IGS Tracking network (one located in Europe and composed by 16 stations around Italy and another one, composed by 9 stations and located in Antarctica) processed respectively for 11 and 3 years, the impact of some data processing strategies has been considered. For both the networks different kind of solutions has been obtained, each one using more refined strategies, considering the ambiguity phase resolution, the possibility to frame the solution using global transformation parameters (provided by JPL), or framing the solution using regional transformation parameters. Furthermore, with the aim to evaluate in polar regions the impact of different tropospheric models in GNSS solutions the Antarctica network has been processed also considering the Vienna Mapping function instead of the Niell MF. The results will be synthetized by two parameters which are accuracy and precision: accuracy has been obtained comparing each solution with the reference ITRF2008 solution (ITRF2008-TRF-IGS.SNX.gz) and precision has been estimated considering the scattering of the solutions (time series) around a straight line obtained by a least square approach.

Le finalità delle reti di stazioni permanenti GNSS, in ambito geodetico, sono: l’erogazione di servizi di posizionamento di precisione in tempo reale, la definizione ed il monitoraggio dei sistemi di riferimento, e il controllo del territorio e delle strutture. In particolare, nei casi di monitoraggio, può essere utile disporre di un dato continuo. Il metodo di calcolo denominato Precise Point Positioning è una valida alternativa all’approccio differenziato e presenta vantaggi in termini di flessibilità, in quanto non impone correlazioni tra stazioni. Esso, così come l’approccio differenziato, consente quindi il calcolo in quasi tempo reale di una qualsiasi rete di stazioni GNSS. In questo lavoro si intende mostrare una procedura, realizzata ad hoc completamente automatica in grado di elaborare grandi reti di stazioni permanenti con ritardo giornaliero o settimanale e precisioni centimetriche sulla singola stima. Tale procedura può essere riassunta in 4 macrofasi: prelievo dati (download dati GNSS e file ancillari), pre-analisi (valutazione consistenza del dataset e individuazione di eventuali errori di natura formale), processamento, post-trattamento (la conversione in diversi sistemi di riferimento o di coordinate, individuazione e successiva rimozione di eventuali outlier nell’ipotesi di comportamento lineare, redazione di risultati grafici). Il campione di dati utilizzati per valutare l’applicabilità della procedura in termini generali consiste in una rete GNSS costituita da 16 stazioni permanenti omogeneamente distribuite sul territorio italiano e zone limitrofe.

Among several instruments and techniques that can be used for monitoring purposes, GNSS technology has undergone a fast evolution and provides a large choice of solutions. Despite the best performances can be achieved by using double frequency geodetic receivers, capable to maintain high precisions even for wide-scale monitoring, in the recent years, several interesting solutions were presented in the low-cost market. In this work, a monitoring system based on a couple of low-cost GNSS receivers has been developed and tested in the field (Ponte Motta in Cavezzo - Modena, Italy). The two receivers were positioned about 50 m from each other and data were acquired at 1-Hz frequency. The position solutions were calculated both in post processing, through two different free and open-source software packages, and using an embedded RTK processor. The analysis aims to assess on one hand the capability of the system to perform the monitoring of slow displacements with the best possible precision, and, on the other hand, the performances of the real-time solutions that can be used for early warning purposes. The precisions evidenced by the tests show that such low-cost instrumentation can be used for many monitoring purposes, especially considering the cost that is about a tenth of geodetic instruments.

WorldView-3 satellite is providing images with an unprecedented combination of high spatial and spectral resolution. The stereo capabilities and the very high resolution of the panchromatic band (0.31 m) have been fostering new applications in urban areas, where the complexity of the morphology requires a higher level of detail. The present technical note aims to test the accuracy of digital elevation models that can be obtained by WorldView-3 stereo-pairs in these particular contexts, with an operational state-of-the-art algorithm. Validation is performed using check points and existing models of the area (from LiDAR data and oblique aerial images). The experiments, conducted over the city of Bologna (Italy) with six images, proved that roof surfaces and open spaces can be reconstructed with an average error of 1–2 pixels, but severe discrepancies frequently occur in narrow roads and urban canyons (up to several metres in average). The level of completeness achievable with only one pair is extremely variable (ranging from 50% to 90%), due to the combined effect of the geometry of acquisition and the specific urban texture. Better results can be obtained by using more than one pair. Furthermore, smaller convergence angles can be beneficial for the reconstruction of specific urban structures, such as soaring towers.

Over the past decade, precise point positioning (PPP) has become a tool widely used in many Global Navigation Satellite Syste applications and the performance levels of the method are often close to those that can be achieved through a differenced approach. The aim of this paper is to analyze the dependency of the precision of a PPP solution output from the GIPSY-OASIS II software on the observing-session durations. In detail, starting from real data acquisitions spanning 24, 12, 6, 3 h, 1 h, and ½ an hour, all processed through the PPP, a continuous function has been defined and the related coefficients have been estimated. The original dataset consists of daily RINEX files provided from 44 sites of both International GNSS Service (IGS) and European Permanent Network (EPN) permanent networks acquired over one year. Each RINEX file was split into several shorter files according to the above listed time spans and processed by using GIPSY-OASIS II together with jet propulsion laboratory (JPL) precise post-processed products. The uncertainty of the proposed function was also estimated and a complete analysis of its compliance with the data sample has been provided. The estimated model is demonstrated to reach a millimeter accuracy level within a statistical confidence level of 99% using the experimental data.