Extending the Applications and Improving the Efficiency of Positioning Through the Exploitation of New GNSS Signals

Lead Research Organisation: Imperial College London
Department Name: Civil & Environmental Engineering

Abstract

Over the past three decades the US GPS (Global Positioning System) has evolved from a system designed to provide metre-level positioning for military applications to one that is used for a diverse range of unforeseen, and mainly civilian, applications. This evolution has been both driven and underpinned by fundamental research, including that carried out at UK universities, especially in the fields of error modeling, receiver design and sensor integration. However, GPS and its current augmentations still cannot satisfy the ever increasing demands for higher performance. For instance there is insufficient coverage in many urban areas, it is not accurate enough for some engineering applications such as the laying of road pavements and receivers cannot reliably access signals indoors.However things are changing rapidly. Over the next few years the current GNSSs (Global Satellite Navigation Systems) are scheduled to evolve into new and enhanced forms. Modernised GPS and GLONASS (Russia's equivalent to GPS) will bring new signals to complement those that we have been using from GPS for the last 30 years. Also we will see the gradual deployment of new GNSSs including Europe's Galileo and China's Compass systems, so leading to at least a tripling of the number of satellite available today by about 2013 - all with signals significantly different from, and more sophisticated than, those used today.These new signals have the potential to extend the applications of GNSS into those areas that GPS alone cannot satisfy. They will also enable the invention of new positioning concepts that will significantly increase the efficiency of positioning for many of today's applications and stimulate new ones, especially those that will develop in conjunction with the anticipated fourth generation communication networks to provide the location based services that will be essential for economic development across the whole world, including the open oceans. This proposal seeks to undertake a number of specific aspects of the research that is necessary to exploit the new signals and to enable these new applications. They include those related to the design of new GNSS sensors, the modeling of various data error sources to improve positioning accuracy, and the integration of GNSSs with each other and with other positioning-related inputs such as inertial sensors, the eLORAN navigation system, and a wide rage of pseudolite and ultra-wide band radio systems. We are also seeking to find new ways to measure the quality of integrated systems so that we can realistically assess their fitness for specific purposes (especially for safety-critical and legally-critical applications). As part of our work we will build an evaluation platform to test our ideas and validate our discoveries.The proposal builds on the unique legacy of the SPACE (Seamless Positioning in All Conditions and Environments) project, which was a successful EPRSC-funded research collaboration framework that brought together the leading academic GNSS research centres in the UK, with many of the most important industrial organisations and user agencies in the field. The project laid the foundation for an effective, long-term virtual academic team with an efficient interface to access industry's needs and experience. The research proposed here will be carried out within a new collaboration framework (based on SPACE) involving four universities (UCL, Imperial, Nottingham and Westminster) and nine industrial partners (EADS Astrium, Ordnance Survey, Leica Geosystems, Air Semiconductors, ST Microsystems, Thales Research and Technology, QinetiQ, Civil Aviation Authority and NSL). The industrial partners have pledged almost 2M of in-kind support and the proposed management structure, led by one of the industrial partners, is carefully designed to foster collaboration and to bring to bear our combined facilities and resources in the most effective manner.
 
Description The objectives of this project were set to undertake a number of specific aspects of the research necessary to exploit the new GNSS signals and enable new applications including those related to the design of new GNSS sensors and modelling of various measurement error sources to improve positioning accuracy, and the integration of GNSSs with each other and with other positioning-related inputs such as ultra-wide band radio systems. We were also seeking to find new ways to measure the quality of integrated systems so that we can realistically assess their fitness for specific purposes (especially for mission-critical applications). The key findings in the particular areas of research are listed below. We specifically considered how the discoveries we make in these could be used to enhance and enable the use of GNSS under the following three specific conditions.

• Precise Point Positioning (PPP), where extensive research was required to model the relevant error sources.

• GNSS based sensor networks, which could result in substantially more sophisticated and higher performance measurement error models.

• GNSS in difficult environments, where research was needed to develop a robust system (including both hardware and software) that could provide metre-level indoor positioning accuracy, and to assess the quality of such a system.



1. The design of new GNSS sensors

The next generation of GNSS receivers should be capable of simultaneously processing existing and new signals. The two challenges faced in the realization of such a receiver are the delivery of the RF signal to the Digital Signal Processor (DSP) with a minimal number of RF/analogue components whilst keeping the noise at a minimum and real-time tracking of the signals transmitted from the satellites with low-power and cost. The key findings in the design of the new GNSS sensors are the:

• development of a configurable GNSS receiver platform for both academic and commercial research into GNSS acquisition and tracking algorithms and implementations

• design of a Radio Frequency (RF) front-end utilising sub-Nyquist sampling to enable a wide range of GNSS signal frequencies to be detected simultaneously while only requiring a single RF signal path and a low component count

• design of a dual-channel RF front-end enabling multipath mitigation experiments using right and left hand polarised antennas

• development of an efficient software acquisition to speed up acquisition 25 times compared with basic FFT approach

• use of an on-board atomic clock as very accurate time reference

• development of an FPGA baseband processor and thereby enabling highly-parallelised DSP along with complex (rather than real-valued only) sampling to extend the signal bandwidth use of time-multiplexing approach to accommodate 96 tracking channels) on the FPGA compared with a 10 channel limitation with the basic solution

• design of a variable correlator spacing tracking channel design for enhanced multipath mitigation and reduction of DLL loop noise

• development of the capability to dynamically change the tracking loop filter parameters and correlator spacing using the connected LINUX workstation

• use of an on-board temperature sensor for the Atomic Clock

• development of a complete digitally configurable GNSS receiver platform capable of real-time GPS L1 L2, Galileo and GLONASS signal tracking.



2. Measurement error modelling

Satellite navigation systems are very complicated. The errors are coming from all segments (control, space and user) and the signal propagation environment. The project investigated in particular the errors in the temporal and spatial reference systems and those related to signal propagation environments including the effects of the ionosphere, troposphere and the vicinity of the receiver/antenna.



(1). Modelling of errors in the temporal and spatial reference systems

The key findings here are that:

• as initially suspected (but not quantified by research) the task of combining ranging observations from different constellations of navigation satellites requires estimates of both space and time frame transformations, and these estimates must be updated periodically because the frame transformations are not constant in time.

• the realisation of any one constellation's spatial reference frame changes, depending upon the tracking station infrastructure used at any one time.

• the US GPS spatial reference frame was degraded in the late 2000s because of the introduction of receiver technology uncalibrated for subtle multipath effects - these instabilities were at the level of 1-3 cm and varied seasonally.

• reference time scales underpinning broadcast clock models are reset by the respective control segments of each system. For GPS this reset typically occurs daily and results in discontinuities in the clock models at the level of a few nano-seconds - mapping into range changes of around a metre.

• while two reference time scales (for respective satellite constellations) do not need to be aligned, it is necessary to estimate an 'inter-constellation bias term' at the observation equation level to absorb the offset.

• when a timescale is reset this aforementioned bias term should be re-estimated from tracking data.

• the effect of this reset on multi-constellation PPP within a Kalman Filter is yet to be assessed but it is worth investigating.



(2). Modelling of the effects of the atmosphere

By slowing and bending the signals from the GNSS satellites, the atmosphere is one of the biggest sources of error in GNSS positioning. Atmospheric effects are conveniently separated into those caused by the upper atmosphere, the ionosphere, and those caused by the lower atmosphere, the troposphere. This project has tackled these two parts of the atmosphere separately, and has made significant progress in mitigating the effects and thereby improving the accuracy and robustness of GNSS positioning.

For precise positioning applications, the best approach to dealing with the troposphere is to estimate the amount by which the signals have been delayed, as part of the positioning solution. This approach further allows the estimated delays to be used for weather forecasting. However, this approach makes no use of the accurate and up-to-date estimates of the state of the troposphere that are produced by, for instance, the UK Meteorological Office. This project has investigated, and demonstrated the benefit of, the use of precise ray tracing techniques to estimate tropospheric delay from the Met Office troposphere products. This improves convergence times in PPP. Furthermore, this project has demonstrated the potential improvement in the GNSS-based estimates of troposphere delays from the availability of multiple constellations of GNSS satellites.

For the ionosphere, this project concentrated on the phenomenon of scintillation, in which rapid variations in ionospheric delay during periods of high ionospheric activity can cause receivers to lose lock on the GNSS signals. This project studied ionospheric scintillation models, and developed an end-to-end approach to simulate scintillation accurately using a Spirent GSS8000 simulator.



(3). Modelling of the errors from the receiver/antenna environment

GNSS signals can be blocked and reflected by buildings and other objects, particularly in urban areas. This produces two phenomena, non-line-of-sight (NLOS) reception and multipath interference, both of which can severely degrade positioning accuracy. Multi-constellation GNSS provides us with the opportunity to reject signals contaminated by these phenomena and compute a position solution from only the good signals. However, the contaminated signals must be identified. Three methods have been developed and tested under the project. The dual-polarization technique detects NLOS signals by comparing the strengths of each signal from the left-hand and right-hand outputs of a dual-polarization antenna. The multi-frequency signal-to-noise comparison technique detects multipath interference by comparing the ratios of the signal strengths on different frequencies with their normal values. Consistency checking exploits the information redundancy in multi-constellation GNSS to identify measurements that are inconsistent with the others; these are likely to be NLOS or multipath-contaminated. A conventional sequential elimination consistency checking algorithm was found to work well in benign environments, but actually degraded the position accuracy in dense urban environments. A new subset-comparison-based consistency checking algorithm has been developed and this works much better.



3. The integration of GNSSs

The novel algorithms and approaches to processing signals from multi-constellation GNSS developed in this project are integrated within a novel core data processing platform (POINT). It enabled the algorithms to be implemented and tested in a consistent, reliable way, using state-of-the-art techniques. Interfaces were developed and implemented to facilitate the integration of the research undertaken in the each of the specific areas. For example, such an interface would allow measurements from high accuracy tropospheric models to be used within the software. A close coupling between the positioning and integrity algorithms formed the basis for a part of the research on integrity.

Based around a Kalman filter, the processing platform has implemented positioning using PPP with measurements from GPS and GLONASS and precise products such as those from the IGS. For GPS positioning, research has also been undertaken with products from CNES for fixing ambiguities with PPP. In addition to PPP research with real observations, the platform has also enabled the development and investigation of algorithms for multi-constellation positioning with simulated data. The Spirent GSS8000 simulator has been used to generate signals from modernised GPS and Galileo. Research has been undertaken that uses the software as a flexible platform for investigating algorithms such as linear combinations of observations to make full use of properties such as high accuracy code measurements.



4. Quality assurance and integrity monitoring

The quality of signals and data processing ultimately affect the performance of positioning and navigation solutions especially for mission (e.g. safety) critical applications. This is measured by the integrity (trust) in the output of the positioning and navigation systems. In order to ensure high performance, the integrity must be monitored over the entire signal/data processing chain. This project focused on high accuracy (centimetre level) GNSS positioning. It employed the carrier phase measurements either in the conventional RTK (cRTK) mode which requires extra data from a dense local network of reference stations to mitigate errors, or in the Precise Point Positioning (PPP) mode, which uses only one receiver with a dedicated communication channel to receive products (error corrections). cRTK is unsuitable for applications in remote areas (including oceans) or expensive to implement/operate such stations. In contrast, Precise Point Positioning (PPP) has the potential for cm-level accuracy using a sparse global network of stations to generate products. In order to support mission critical (e.g. safety) applications, this project has developed a comprehensive integrity monitoring process. The key findings are the development of:

• a single receiver measurements based data processing methodology to significantly reduce residual errors.

• high accuracy PPP algorithms able to resolve and fix the integer ambiguity of GPS carrier phase observations.

• multiple-constellation PPP algorithms enabling fast ambiguity resolution of GPS carrier phase observations, aided by GLONASS measurements.

• an enhanced ambiguity validation method through the discovery of the distribution of the Ratio test statistic as doubly non-central F distribution.

• a reliable confidence level based Ratio test methodology

• new multiple-Constellation GNSS positioning algorithms

• of new multiple-Constellation GNSS integrity monitoring algorithms
Exploitation Route Non-academic context can use the findings from this project in different ways including adaptation the concept/methodologies/algorithms already in the public domain, purchases of IPs, and collaboration to develop the findings further.

The novel receivers, designed during the period of this project, have the potential to be commercialised. The FPGA based GNSS platforms can be incorporated in the development of new technologies for the GNSS. Making use of the widely accepted VHDL programming language the insertion of new algorithms and techniques for various stages of the signal acquisition and processing is possible. The FPGA based GNSS platforms can be used by R&D departments to evaluate and demonstrate their new products. There is also the possibility to use the dual RF channel receiver for educational purposes. The GNSS platform has been used in several occasions by the University of Westminster for their open events and to support the syllabuses for telecommunication courses.

The work on tropospheric delay estimation from Met Office products has the potential to improve the accuracy of real-time positioning, particularly for receivers using the Precise Point Positioning technique. It is envisaged that current real-time augmentation networks could be used to estimate localized improvements to the Met Office data for broadcast to local users. This technique would then allow receivers to apply this accurate external model to correct for tropospheric delay and therefore to achieve higher accuracy and quicker convergence in a PPP solution.

The scintillation knowledge gained from the ionosphere work could pave the way for the next generation ionosphere mitigation receiver. Based on external ionosphere monitoring information, the current receiver positioning technique can be improved, in particular against the phenomenon of ionosphere scintillation, which is a severe problem in equatorial and high latitude regions. The scintillation mitigation techniques will improve the GNSS application usage in these affected regions, and benefit applications such as precision agriculture, surveying, land management and off-shore operations.

The dual-polarization, multi-frequency signal-to-noise comparison and consistency checking techniques for detecting NLOS reception and multipath interference could all be implemented in commercial GNSS user equipment with further development. This would improve the accuracy and reliability of GNSS positioning in urban areas, supporting navigation, surveying and a wide-range of location-based services. Further improvements could be achieved by integrating GNSS with dead reckoning and dissimilar positioning technologies.

The research on multi-constellation processing remains a subject of current research and the optimum combination of signals from different systems is critical to obtaining the best results from GNSS receivers. This research therefore has potential to be exploited by receiver manufacturers.

The processing filter uses a flexible plug-and-play approach to sensor integration and novel approaches for integrating GPS and low cost INS data to provide accurate roll, pitch and yaw information on a platform. This research has the potential to be used in many sectors, for instance in the automotive/motor sport sector where real-time velocity and attitude are required in a highly dynamic environment.

The research achievements in integrity monitoring have attracted the attention of the major companies in this area including Novatel, Leica Geosystems and Veripos. Companies can adopt the algorithms to new products and the development of new services, or upgrade existing products and services. The potential collaboration with Veripos is under discussion to improve the company's product and service especially the integrity monitoring for PPP.

The industrial collaborators on the project have and will continue to benefit from the findings from this project in the enhancement of their current products and services, and the development of new ones.
The methodologies and algorithms developed have been extensively tested in this project with simulated data with Sprient GNSS simulator, and real data from reference stations around the world. This research can be put to use through knowledge transfer, consultancy and direct research collaboration with companies. In addition, the PhD students and postdoctoral researchers trained in this project can directly work for commercial companies. Four postdoctoral researchers and three PhD students who worked on this project have been employed by leading companies in GNSS including Intel and Novatel. The particular findings can be exploited by individually or combinations thereof.

The dual RF end receiver design can be used in research applications, such as multipath mitigation, by connecting the RF channels to left-hand and right-hand polarised antennae and satellite clock modelling by capturing different GNSS signals from each front-end to perform a differential analysis of the clock biases from both systems. The effect of scintillation on signal tracking can also be analysed with the receiver since such analysis requires input spectral diversity with a high update rate and a stable reference clock. The receiver makes it possible to deploy a diverse range of applications involving rapid development, real-time prototyping and assessment of new architectures, circuits and systems for GNSS receivers.

The tropospheric work formed the basis of a proposal to ESA to develop new troposphere models, and is being further investigated with the use of a national network of continuously operating GNSS receivers.

The findings of the ionospheric work have contributed to an FP7 project, 'Countering GNSS High Accuracy Applications Limitations Due To Ionospheric Disturbances In Brazil' (CALIBRA). In the CALIBRA project, the ionosphere scintillation mitigation methods were studied, based on the scintillation knowledge learnt from this project.

The dual-polarization, multi-frequency signal-to-noise comparison and consistency checking techniques for detecting NLOS reception and multipath interference could all be implemented in commercial GNSS user equipment with further development. We have collaborated with Novatel (part of the same company as Leica) on testing the dual-polarization technique and with ST Microelectronics on testing the consistency checking technique.

The data processing platform has been used extensively within the project, including as a testing ground for algorithms relating to tropospheric delay, Precise Point Positioning, and integrity. Due to its flexibility, it is regularly used within the Nottingham Geospatial Institute (NGI) as a processing tool for other research projects, including an EC FP7 project on using vision sensors to aid GNSS, and as a real-time processing engine for the NGI's 'foot tracker' pedestrian navigation system.

The integrity monitoring algorithms developed for carrier phased based integrity monitoring can be used with new signals from multiple constellations in both RTK and PPP modes. It provides not only the techniques to support mission critical applications but also an indicator to support certification and regulation in the use of GNSS. This work will be exploited as a part of a new COST ACTION project involving Imperial College London, on the standardisation and certification of GNSS-based Intelligent Transport Systems (ITS).

It is important to note that this project was undertaken in collaboration with the leading GNSS companies in the UK. It is expected that, over the next few years, these companies will in conjunction with the relevant business units at the Universities, develop and implement a plan to commercially exploit the findings.
Sectors Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Electronics,Energy,Environment,Security and Diplomacy,Transport

URL http://www.insight-gnss.org/
 
Description The industrial collaborators on the project are benefiting from the findings from this project in the enhancement of their products and services, and the development of new ones.
First Year Of Impact 2013
Sector Aerospace, Defence and Marine,Agriculture, Food and Drink,Construction,Digital/Communication/Information Technologies (including Software),Energy,Environment,Transport
Impact Types Societal,Economic

 
Description iNSIGHT 
Organisation Air Semiconductor Ltd
Country United Kingdom 
Sector Private 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation Department of Transport
Department Civil Aviation Authority (CAA)
Country United Kingdom 
Sector Public 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation EADS Astrium
Country France 
Sector Private 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation Leica UK Ltd
Country United Kingdom 
Sector Private 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation Nottingham Scientific
Country United Kingdom 
Sector Private 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation Ordnance Survey
Country United Kingdom 
Sector Public 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation ST Microelectronics
Country Switzerland 
Sector Private 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation Thales Group
Department Thales Research & Technology (Uk) Ltd
Country United Kingdom 
Sector Private 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation University College London
Department Department of Civil, Environmental and Geomatic Engineering
Country United Kingdom 
Sector Academic/University 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation University of Nottingham
Department Nottingham Geospatial Institute
Country United Kingdom 
Sector Academic/University 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009
 
Description iNSIGHT 
Organisation University of Westminster
Country United Kingdom 
Sector Academic/University 
PI Contribution Imperial College London developed the integrity monitoring algorithms
Collaborator Contribution The industrial partners contributed to overall direction of research, UCL on multipath modelling, Westmister on receiver design and Nottingham on sensor integration.
Impact The publications are listed in the relevant section
Start Year 2009